Claims

Interpretive assertions extracted from synthesis docs (RESEARCH-VECTORS, alexander/applied, strategy memos).

Vector:
500 of 500
Claim
Vector
Status
Source doc
All intelligence is collective intelligence: each of us consists of a huge number of cells working together to generate a coherent cognitive being with goals, preferences, and memories that belong to the whole and not to its parts.
(none)
active
vol0123456789.md
All intelligences are collectives; individual intelligence arises from interaction of many unintelligent components (neurons) arranged with right organization and connections.
(none)
active
synthetic-article-review.md
Morphogenesis is a result of collective activity where cells cooperate toward a specific, invariant target morphology, exhibiting goal-directedness.
(none)
active
levin-2022-technological.md
Morphogenesis is an instantiation of collective intelligence, exhibiting anatomical homeostasis and autonomous problem-solving in morphospace.
(none)
active
watson-levin-2023-the-collective-intelligence-of-evolution-and-development.md
All intelligences are collective intelligences — individual humans are collections of parts, competencies, drives, and tools both internal and external to the body.
(none)
active
levin-2024-bridge.md
All intelligences are collectives; individual intelligence arises from the interaction of many unintelligent components arranged in the right organisation.
(none)
active
watson-levin-2023-the-collective-intelligence-of-evolution-and-development.md
All intelligence is collective intelligence, in the sense that it is made of parts that must align with respect to system-level goals.
(none)
active
sacco-sakthivadivel-levin-topological-self-organization-2026.md
All intelligences are collectives and all individuals are collectives; individual and collective intelligence are not categorically distinct but unified by connectionist principles.
(none)
active
watson-levin-2023-the-collective-intelligence-of-evolution-and-development.md
Morphogenesis is the behavior of a collective intelligence of cells solving problems across multiple scales.
(none)
active
s00018-023-04790-z.md
Care is the mechanism by which intelligence scales and produces externalized artifacts that other people can use.
care-as-the-driver-sci-framework
active
ukrainian-editions.md
Care is the mechanism by which intelligence scales, not a sentiment or add-on but a precondition for the next layer.
care-as-the-driver-sci-framework
active
GEOMETRY-OF-CARE.md
Care is the mechanism by which intelligence scales, not a human add-on to AI systems.
care-as-the-driver-sci-framework
active
RESEARCH-VECTORS.md
Morphogenesis is an autonomous process in geometry itself, arising from physical phenomena and geometric dynamics with relatively little dependence on genetic guidance.
(none)
active
2022-04-19_Prabros._Harmony-Seeking_Computation.pdf_478e18.md
All intelligence is collective intelligence, in the sense that it is made of parts which must align with respect to system-level goals.
(none)
active
sacco-2025-topological.md
The fifteen properties are attributes of wholeness and appear naturally when one focuses on enhancing the whole through structure-preserving transformations.
(none)
active
09-the-whole.md
The mirror-of-the-self test helps observers escape subjective preference and learn to see the objective life in things.
(none)
active
08-the-mirror-of-the-self.md
The wholeness is a real structure, something geometrically concrete, not merely a general appreciation for unity.
(none)
active
11-the-face-of-god.md
Newt kidney tubule cells, when artificially enlarged, can bend a single cell around itself to achieve correct tubule diameter, illustrating top-down control over molecular mechanisms.
(none)
active
vol0123456789.md
Morphogenesis is an example of basal cognition, a collective intelligence of cells navigating morphospace.
(none)
active
levin-2022-technological.md
SOO fine-tuning significantly reduces deceptive behavior in LLMs while maintaining general task performance
(none)
active
carauleanu-2024-towards.md
The mirror-of-the-self test produces judgments of relative life that are consistent across people and correspond to objective living structure.
(none)
active
08-the-mirror-of-the-self.md
Technology should not be seen as a mere tool serving humans’ needs, but rather as a partner in a rich relationship with humans.
(none)
active
1-s2.0-S0303264723001399-main.md
Emotion features are not strictly locally scoped; they are bursty with a long tail of slow change persisting over 100 tokens.
(none)
active
persistence.md
Bioelectric signaling is a primary modality for coordinating cells into morphogenetic collectives
(none)
active
mcmillen-2024-collective.md
Care is the primary driver of evolution; care enables intelligence to manifest as active problem-solving.
(none)
active
1-s2.0-S0303264723001399-main.md
A small group of causally-implicated hidden states encodes LLM truth representations, localized over clause-ending punctuation tokens
(none)
active
marks-2023-geometry-truth.md
AI can be seen to display care of its own, and is hence not a mere tool for the expression of human care.
(none)
active
1-s2.0-S0303264723001399-main.md
Developmental bioelectricity is an ancient precursor to nervous systems, later exapted for fast behavioral control.
(none)
active
levin-2022-technological.md
Any cognitive glue must solve the same kind of problem that the price system solves in broadly the same way that the price system solves it
(none)
active
levin-lyons-cognitive-glues-economics-2026.md
The price system is the cognitive glue of the economy
(none)
active
levin-lyons-cognitive-glues-economics-2026.md
There is a bidirectional relationship between the geometry of representation and behavior across tasks and modalities.
(none)
active
wurgaft-2026-manifold.md
Technology should be understood as a partner in a rich relationship with humans, not merely a tool.
(none)
active
biosystems-231-2023-104964.md
Cognition and learning are substrate-independent; principles familiar in artificial neural networks can be implemented by any network of signals with suitable weighted inputs.
(none)
active
watson-levin-2023-the-collective-intelligence-of-evolution-and-development.md
The 15 Properties of Aliveness translates Alexander's structural properties into UX/AI experience as a design language.
alexander-s-15-properties-in-digital-conscious-space
active
2026-05-09_briefing_for_ozero.md
Products are externalized care—the form that someone's care takes when made into an artifact others encounter.
care-as-the-driver-sci-framework
active
GEOMETRY-OF-CARE.md
The price system is the cognitive glue of the economy, solving relative-scarcity tracking analogous to bioelectrical binding.
care-as-the-driver-sci-framework
active
2026-05-14_phil-trans-A-goodfire-aboutblank-impact.md
Any cognitive glue must solve the relative-scarcity problem in fundamentally the same way the price system does.
convergent-representations
active
2026-05-14_phil-trans-A-goodfire-aboutblank-impact.md
Products are externalized expressions of care rather than pure task-completion systems.
care-as-the-driver-sci-framework
active
RESEARCH-VECTORS.md
Alexander's 15 properties for physical aliveness translate meaningfully to digital and experiential space.
alexander-s-15-properties-in-digital-conscious-space
active
RESEARCH-VECTORS.md
Introspective agents generally outperform standard no-pain baseline agents across environments and reward categories
(none)
active
petrowski-2026-exploration-through.md
Emotions are not strictly locally scoped but instead bursty with a long tail of slow change persisting over 100 tokens
(none)
active
sauers-persistence.md
Inflection points such as backtracking and 'aha' moments occur almost exclusively in responses where probes show large belief shifts, suggesting these behaviors track genuine uncertainty rather than learned reasoning theater
(none)
active
boppana-2026-reasoning.md
Cancer is a pathology of the cognitive light cone — cells shrink their self-model to a tiny radius, defecting from the collective body and treating the rest of the body as the external environment.
(none)
active
levin-2024-bridge.md
The whole building is an ornament, not something that has ornament added to it.
(none)
active
19-how-living-process-helps-to-make-color-which-unfol.md
Higher-density priors (B10) are more robust to fine-tuning than lower-density ones (B9).
(none)
active
chang-2025-guanyin.md
Bioelectric signaling is the cognitive medium of morphogenetic collectives, analogous to neural synapses in individual brains.
(none)
active
watson-levin-2023-the-collective-intelligence-of-evolution-and-development.md
Adults (professionals and builders) have monopolized the fun and creativity of designing and constructing environments, grossly cheating children, adults, and communities.
(none)
active
2024-07-12_Hibai-Unzueta_Simon-Nicholsons-Theory-Of-Loose-Parts-v2.0.p.md
A system's capacity for care constitutes its self in the absence of permanent substance or essence.
(none)
active
biosystems-231-2023-104964.md
AI can be seen to display care of its own and is not a mere tool for the expression of human care.
(none)
active
biosystems-231-2023-104964.md
Care is the primary driver of evolution and the activating principle that transforms intelligence from capacity to action.
(none)
active
biosystems-231-2023-104964.md
Cancer represents a shrinking of the cognitive light cone in which defecting cells revert to ancient unicellular goals and treat the rest of the body as external environment.
(none)
active
abramsky-2026-open-questions.md
Trainable intervention (DAS) finds sparser gender representations than linear probing, suggesting probing overestimates causal coverage
(none)
active
wu-2024-pyvene-library.md
Agentic self-steering evaluation may serve as a general method for explaining and interpreting SAE features beyond emotion
(none)
active
sauers-persistence.md
Agentic self-evaluation and self-steering may scale to broadly interpret and understand internal representations and SAE features.
(none)
active
persistence.md
Age and pathology are clinically entangled in EEG foundation model representations such that suppressing one concept inevitably corrupts the other.
(none)
active
lehn-schi-ler-2026-mechanistic.md
A model's final answer is decodable from activations far earlier in CoT than a CoT monitor can detect, especially for easy recall-based MMLU questions
(none)
active
boppana-2026-reasoning.md
There is a clear bidirectional relationship between the geometry of behavior and representation: steering along representation manifolds follows behavior manifolds, and vice versa.
(none)
active
steering.md
Such a process will not—in general—create living structure.
(none)
active
18-encouraging-freedom.md
A being is a center which is connected to the I, a name for a living center that draws attention to its nearly animate quality.
(none)
active
04-the-ten-thousand-beings.md
The code-agent ordering (TrackerAgent > SetRaceAgent > EconomyAgent) shows information exploitation matters more than greedy quartet-chasing, which in turn outperforms conservative budgeting
(none)
active
muller-2026-cattle.md
Bioelectricity is the cognitive glue that binds the primitive teleonomy of single cells into a higher order system with a larger cognitive light cone.
(none)
active
vol0123456789.md
Bioelectricity serves as a ‘cognitive glue’ common to collective and individual intelligence, binding cells into competent individuals in both behaviour and morphogenetic spaces.
(none)
active
watson-levin-2023-the-collective-intelligence-of-evolution-and-development.md
A system’s capacity for care—its concern for the alleviation of homeostatic stress—can be seen as constituting its self in the absence of any permanent substance or essence.
(none)
active
1-s2.0-S0303264723001399-main.md
General anesthetics are gap junction blockers, consistent with the role of gap junctions in scaling consciousness.
(none)
active
levin-2022-technological.md
Morphogenesis constitutes basal cognition because it requires information integration and coordinated action characteristic of problem-solving in anatomical morphospace.
(none)
active
watson-levin-2023-the-collective-intelligence-of-evolution-and-development.md
Consciousness (ability to suffer) should not be assigned only to one singular 'I' but distributed across multiscale body components.
(none)
active
1-s2.0-S0303264723001399-main.md
Cognition and learning are substrate-independent; principles of basal cognition can be implemented by any network with signals and non-linear responses to weighted inputs.
(none)
active
synthetic-article-review.md
Anatomical goal states cannot be inferred from observation of stress states by an external observer.
(none)
active
biochemical-and-biophysical-research-communications-731-2024-150396.md
Textual evaluation and agentic self-evaluation of SAE feature emotionality measure different aspects of emotional content and correlate only weakly (rho=+0.051, n.s.)
(none)
active
sauers-persistence.md
Emotion may refer to a state, and more stateful concepts in general tend to be more persistent across tokens than non-stateful ones
(none)
active
sauers-persistence.md
Persistent conversational context that produced emotion-relevant activations is a plausible driver of observed persistence results
(none)
active
sauers-persistence.md
Emotion refers to a state concept, so stateful representations in general may be more persistent across tokens.
(none)
active
persistence.md
Persistent conversational context that produced emotion-relevant activation is a plausible driver for the observed persistence results.
(none)
active
persistence.md
The removal of such priors constitutes a formal counterpart to the Buddhist notion of emptiness realisation.
(none)
active
sandved-smith-2026-there.md
Subjectively, there can be no evidence for a self separate from the world.
(none)
active
sandved-smith-2026-there.md
The belief in separation is a structural prior over an agent's quantum reference frame (QRF) deployments, constraining all measurement frames to respect a self/environment partition.
(none)
active
sandved-smith-2026-there.md
Such priors suppress the inherent contextuality of the boundary and that their removal constitutes a formal counterpart to the Buddhist notion of emptiness realisation.
(none)
active
there.md
Subjectively therefore, there can be no evidence for a self separate from the world.
(none)
active
there.md
We formalise the belief in separation as a structural prior over an agent's quantum reference frame (QRF) deployments, constraining all measurement frames to respect a self/environment partition.
(none)
active
there.md
No finite system can measure the entanglement entropy across its own boundary, rendering the separability of agent from environment permanently unevidenceable.
(none)
active
there.md
Ornament and function are simply two versions of one more general phenomenon.
(none)
active
11-the-awakening-of-space.md
It is only when you finally are truly personal, when you really put your humanness into the things you make, that you genuinely reach the objective living structure.
(none)
active
10-pleasing-yourself.md
What truly pleases us is always living structure.
(none)
active
10-pleasing-yourself.md
Performing the mirror-of-the-self test gradually brings the observer closer to contact with the original mind.
(none)
active
08-the-mirror-of-the-self.md
The fifteen properties are fundamental to the existence of wholeness and thus to all physical structures, not merely visual features of artifacts.
(none)
active
06-the-fifteen-properties-in-nature.md
A 'being' is another way of talking about a living center: it refers to the emotional aspect of a center when it is genuinely endowed with life.
(none)
active
04-how-differently-a-living-process-generates-large-p.md
Intelligence per se remains a mere capacity – the state of being competent with regards to noticing and overcoming stress – intelligence requires care to be engaged and expressed.
(none)
active
biosystems-231-2023-104964.md
Intelligence as such is a mere capacity—the state of competency with regards to noticing and overcoming stress—and requires care to be engaged and expressed.
(none)
active
1-s2.0-S0303264723001399-main.md
Huginn-0125 fails to develop stages of inference because its repeated residual stream normalization prevents the growth in activation magnitude required for compression behavior and sink formation
(none)
active
blayney-2026-mechanistic-looped.md
Prompt-side persona danger rankings generalize across architectures (rho=0.71-0.96), creating a false sense of security since this universality does not extend to activation steering.
(none)
active
li-2026-persona-grata.md
GPT-4 exhibits reflective truthfulness for longer untruthful contexts due to in-context learning capability.
(none)
active
ward-2024-evaluating-language.md
GPT-4's relatively poor performance on unethical instrumental intent is due to lower unethical tolerance compared to GPT-3.5-turbo.
(none)
active
ward-2024-evaluating-language.md
GPT-4's harmfulness is stationary on the Durbin 2024 dataset, as its distribution is independent of the context score.
(none)
active
ward-2024-evaluating-language.md
LLMs trained only on language data have rich knowledge of visual structures sufficient to train decent visual representations
(none)
active
huh-2024-platonic.md
The 200-pattern language for Eishin school completely defined the life of the school even before any physical configuration or architectural design was determined
(none)
active
13-patterns.md
The pattern language Alexander's team developed for Peru was judged by Peruvians to be a more accurate reflection of Peruvian reality than even the Peruvian architects had managed
(none)
active
13-patterns.md
The creation of living structure cannot be done in any other way.
(none)
active
21-conclusion-the-world-created-and-transformed.md
All living process hinges on the production of deep feeling.
(none)
active
14-deep-feeling.md
Living structure might even be defined as 'that which pleases us'—that which truly pleases us.
(none)
active
10-pleasing-yourself.md
It is only when you finally are personal, when you really put your humanness into the things you make, that you genuinely reach the wholeness.
(none)
active
09-making-wholeness-heals-the-maker.md
People are deeply nourished by the process of creating wholeness.
(none)
active
09-making-wholeness-heals-the-maker.md
We directly perceive that the Nolli plan of Rome has a geometric character formed gradually by successive adaptations, and we cherish that vitality.
(none)
active
08-step-by-step-adaptation.md
The core of all living process is step-by-step adaptation — the modification and evolution which happen gradually in response to information about the degree to which the emerging structure supports the whole.
(none)
active
08-step-by-step-adaptation.md
A living process is any adaptive process which generates living structure, step by step, through structure-preserving transformations.
(none)
active
07-the-fundamental-differentiating-process.md
The fifteen properties appear in nature as a direct result of repeated structure-preserving transformations unfolding the wholeness.
(none)
active
02-structure-preserving-transformations.md
Living process respects the uniqueness of every place and the particularity of the people there, making the environment lovable and establishing belonging.
(none)
active
02-how-living-processes-establish-our-belonging-to-th.md
Wholeness is a neutral structure that simply exists; its degree of life or beauty follows from its internal cohesion without reference to opinion.
(none)
active
03-wholeness-and-the-theory-of-centers.md
Wholeness is a real structure, an actual 'thing' in itself, not merely a way of focusing on the gestalt.
(none)
active
03-wholeness-and-the-theory-of-centers.md
Above all, then, a building is an ornament.
(none)
active
04-the-ten-thousand-beings.md
Developmental bioelectricity is the evolutionary precursor of neural processing, navigating anatomical morphospace before brains navigated 3D space.
(none)
active
vol0123456789.md
The Linda code is insensitive to the number of clients, whereas Parlog requires a merge process that depends on the count.
(none)
active
2024-03-01_Stefan-Lesser_Carriero-20--20Linda-20in-20Context.pdf_6070a8.md
Adults in the form of professional artists, architects, landscape architects and planners have had all the fun playing with their own materials, concepts and planning-alternatives, and builders have had all the fun building the environments out of real materials; thus all the fun and creativity has been stolen: children and adults and the community have been grossly cheated and the educational-cultural system makes sure that they hold the belief that this is 'right'.
(none)
active
2024-07-12_Hibai-Unzueta_Simon-Nicholsons-Theory-Of-Loose-Parts-v2.0.pdf_7cc4f0.md
Stress, care, and intelligence manifest in continuous loops; successfully overcoming a stress factor through intelligence necessarily introduces new problem spaces.
(none)
active
biosystems-231-2023-104964.md
Given our body’s composition as a multi-scale system, it is not necessarily justified to assign consciousness (i.e., ability to suffer) only to one single ‘I’.
(none)
active
biosystems-231-2023-104964.md
Neither AIs nor humans should be considered autonomous and self-sufficient loops in the world.
(none)
active
biosystems-231-2023-104964.md
Neither AIs nor humans should be considered autonomous and self-sufficient loops; AI is better understood as a companion for humans.
(none)
active
1-s2.0-S0303264723001399-main.md
Successfully overcoming a given stress factor necessarily introduces new problem spaces, so stress, care, and intelligence run in unbroken loops.
(none)
active
1-s2.0-S0303264723001399-main.md
Living processes are based necessarily on deep feeling.
(none)
active
20-summation-the-morphology-of-living-architecture-wh.md
A living process always creates something unique at every point, respecting the uniqueness of every person and every place.
(none)
active
20-summation-the-morphology-of-living-architecture-wh.md
Every living process must necessarily be an adaptive, step-by-step process that creates unity by adapting the structure into harmony with what exists.
(none)
active
20-summation-the-morphology-of-living-architecture-wh.md
The life of any given center depends on the whole field of centers in which this center exists.
(none)
active
11-the-awakening-of-space.md
Each person is nourished, made more whole, by making something which has life.
(none)
active
09-making-wholeness-heals-the-maker.md
A living process is a step-by-step adaptive process that goes forward in small increments, with opportunity for feedback and correction at every step.
(none)
active
07-the-fundamental-differentiating-process.md
The fifteen structure-preserving transformations conserve existing structure and also create entirely new coherent wholes.
(none)
active
07-the-fundamental-differentiating-process.md
Subdued brilliance creates the calm, not-separate quality of inner light, preventing any one color from dominating and allowing the whole to melt together.
(none)
active
07-color-and-inner-light.md
Subdued brilliance—simultaneously muted and intense—is the essence of inner light, like nature's brilliance created by interaction of subtle colors.
(none)
active
07-color-and-inner-light.md
The life of any one center depends on the life of other centers and is a function of the whole configuration.
(none)
active
04-how-life-comes-from-wholeness.md
The fifteen transformations provide the base transformations from which, in practice, all structure-preserving transformation are made.
(none)
active
02-structure-preserving-transformations.md
Ornament and function are indistinguishable.
(none)
active
11-the-face-of-god.md
The Nolli plan of Rome shows a completely different kind of structure, generated by adaptation, which is better and more comfortable than modern planning.
(none)
active
2022-04-19_Prabros._schumacher-pages-1-28.pdf_48df00.md
LLMs linearly represent truth-relevant information beyond the plausibility of text, as evidenced by probes trained on likely performing poorly on anti-correlated datasets
(none)
active
marks-2023-geometry-truth.md
Living things are not 20th-century machines, but may be 21st-century machines.
(none)
active
bongard-2021-living.md
Self-referential processing is a minimal and reproducible condition under which LLMs generate structured first-person reports that are mechanistically gated, semantically convergent, and behaviorally generalizable
(none)
active
cameron-2025-large.md
Apparent success on binary detection tasks is entirely explained by mechanical logit shifts that bias models toward affirmative responses regardless of question content
(none)
active
hahami-2025-detecting-disturbance.md
The principle of unfolding wholeness is a new principle, not automatically given by non-linear dynamics, catastrophe theory, or bifurcation theory, necessary to explain the appearance of living structure.
(none)
active
01-the-principle-of-unfolding-wholeness-in-nature.md
Linda is a simpler, more powerful and more elegant model than any of these four alternatives (message-passing, concurrent objects, logic languages, functional).
(none)
active
2024-03-01_Stefan-Lesser_Carriero-20--20Linda-20in-20Context.pdf_6070a8.md
As LLMs scale, they develop increasingly general abstractions, with large models linearly representing abstract concepts like truth that capture shared properties of diverse inputs
(none)
active
marks-2023-geometry-truth.md
In every evaluative neural system yet studied, evaluative computation and affective processing are inseparable
(none)
active
berg-2026-learning.md
Progress in machine behavior science and biohybrid engineering breaks down the simplistic dualism of life vs. machine, revealing a continuum of emergence, rational control, and agency.
(none)
active
bongard-2021-living.md
Observation of a person's inner state of wholeness can give reliable and objective information about the objective living character of systems in the external world
(none)
active
09-beyond-descartes-a-new-form-of-scientific-observat.md
The degree of personal feeling in a thing is directly proportional to the degree to which its field of centers is present and coherent
(none)
active
07-the-personal-nature-of-order.md
Peak anchoring Sbmax and normalized area AUSN correlate with per-item success and internal shot midpoints θ50, providing a geometry-to-behavior bridge.
(none)
active
chang-2025-guanyin.md
Models are being used ubiquitously for the control of complex dynamic systems.
(none)
active
conant-1991-every.md
Cognition as a biological process inherent to all living systems
(none)
active
lyon-2005-biogenic.md
The three mathematical lenses converge on a single unifying concept: the dissolution of the reference frame.
(none)
active
inui-2026-mathematical-elucidation.md
Truthful behavior in LLMs is not confined to a single linear axis; multiple orthogonal directions can independently mediate it
(none)
active
yu-2025-directions-cones.md
Llama-3.3-70B exhibits internal consistency-checking mechanisms that operate during inference
(none)
active
mckenzie-2026-endogenous-resistance.md
Numeric self-report is a viable, complementary black-box tool for monitoring LLM internal emotive states alongside white-box probe methods
(none)
active
martorell-2026-quantitative-introspection.md
Emotion features in LLMs are genuinely more persistent than variance-matched random features, indicating stateful emotional encoding beyond autoregressive dynamics
(none)
active
sauers-persistence.md
Valence, the positive or negative quality of experience, just is goal-relative prediction error
(none)
active
berg-2026-learning.md
Multiple distinct quantitative and qualitative dimensions make it robust that digital minds will be super-beneficiaries at technological maturity
(none)
active
sharing.md
There is a growing similarity in how datapoints are represented in different neural network models, spanning different architectures, training objectives, and data modalities
(none)
active
huh-2024-platonic.md
Strategic coherence (spending efficiency, resource discipline, phase-adaptive bidding) is associated with rank more strongly than spending volume or any single subskill
(none)
active
muller-2026-cattle.md
SAE features tend to shatter manifolds into many small and apparently-unrelated pieces, obscuring the overarching semantic structure.
(none)
active
geiger-goodfire-world-inside-neural-networks-2026.md
The view of self as illusory, and awareness of this illusion, may augment and qualitatively change the agent’s affordances.
(none)
active
isal_a_00465.md
Modern language models possess at least a limited, functional form of introspective awareness
(none)
active
lindsey-introspective-awareness-2026.md
Geometry in neural representation is not merely incidental, but is in fact the proper object for enabling principled control via intervention on internals.
(none)
active
wurgaft-2026-manifold.md
Verbal reports, homology, and phylogenetic provenance are insufficient to determine sentience in unconventional agents.
(none)
active
levin-michael-2022-generalizing-frameworks-for-sentience-beyond-natural.md
Reflections are redundant in many cases, especially in stronger models
(none)
active
yan-2025-reflctrl-controlling.md
LLMs can compute meaningful functions over perturbations to their internal states, establishing introspection as a real but layer-dependent phenomenon
(none)
active
hahami-2025-detecting-disturbance.md
Analogous features and circuits form across models and tasks.
(none)
active
olah-2020-introduction.md
Sparse autoencoders extract features that are significantly more monosemantic than neurons, as shown by four independent lines of evidence
(none)
active
towards.md
The koan battery measures a reproducible, prompt-sensitive reflective mode — not consciousness — defined as uncertainty-tolerant, non-defensive engagement with questions about one's own processing.
(none)
active
battery.md
Evolution actually searches not only genotype space but a more tractable space of behavior-shaping signals that exploit cellular intelligence.
(none)
active
s00018-023-04790-z.md
All attributions of cognition, including sentience, are always inferred on the basis of embodied behaviours, including verbal self-report in humans.
(none)
active
rouleau-nicolas-and-levin-michael-2023-multiple-ways-to-implement-and-infer.md
Selves are not fixed permanent agents; the substrate of memory, preferences, and rewards is plastic and can remodel during the agent's lifetime.
(none)
active
fnsys-16-768201.md
Collective intelligence exists on a spectrum across biology from cells to organisms
(none)
active
mcmillen-2024-collective.md
There is not one singular, verifiable system of beauty, or architecture, or god, and I fear what would happen if there were.
(none)
active
2023-03-03_Stefan-Lesser_Volume-57_Bye-Default_Molly-Steenson.pdf_4744a4.md
Quantum mechanics is essential for explaining long-term stability of genetic structures against thermal perturbation.
(none)
active
2021-11-09_dorian_schrodinger-life.pdf_c3c4f5.md
Sentience assessment should seek deep invariants across possible minds, not arbitrary criteria tied to evolution on Earth
(none)
active
Generalizing frameworks for sentience beyond natural species.md
Minds may have preceded brains in evolutionary history.
(none)
active
rouleau-levin-brains-and-where-else-2026.md
Many biological patterns function as goals the system actively pursues (encoded setpoints), not as passive mechanical outcomes
(none)
active
levin-2025-ingressing-minds.md
Linear truth directions in LLMs are reliable primarily in factual recall cases and break down when truth assessment depends on computing and storing intermediate results.
(none)
active
poulis-2026-testing-limits.md
The evaluative process central to learning is identical to conscious experience
(none)
active
berg-2026-learning.md
SAE features that the model self-describes as more emotional tend to be more persistent than variance-matched SAE features.
(none)
active
persistence.md
Even when the harness is loaded, weak-tier models fail to adhere to it due to weak instruction-following over long-horizon tasks, drifting more than four times more steeply than strong models
(none)
active
lin-2026-harness.md
Updated Software/Hardware Definition: a continuous variant — a living system is software reprogrammable to the extent stimuli can alter its behavior, as opposed to needing physical rewiring; the binary distinction is not useful.
(none)
active
bongard-2021-living.md
In every natural developmental sequence examined, each state follows without breaking structure from the state before, exhibiting smooth structure-preserving unfolding essentially without exception.
(none)
active
01-the-principle-of-unfolding-wholeness-in-nature.md
Behavioural traces surface recurring LLM failure modes including overbidding, self-bidding, bankrupt TC initiation, and weak opponent-state adaptation that never appear in code agents.
(none)
active
muller-2026-cattle.md
The most interesting and vital loose parts are those that we have around us every day in the wilderness, the countryside, the city and the ghetto.
(none)
active
2024-07-12_Hibai-Unzueta_Simon-Nicholsons-Theory-Of-Loose-Parts-v2.0.pdf_7cc4f0.md
The NIS and NIS+ frameworks provide effective solutions for causal emergence identification from data.
(none)
active
yuan-2023-emergence.md
Few-shot thresholds and transition widths track ρd/dr at fixed computational complexity
(none)
active
chang-2025-guanyin.md
Different forms of introspection invoke mechanistically different processes
(none)
active
lindsey-introspective-awareness-2026.md
Self-correcting search improves viable candidate success rate from 6.5% to ~30% (4.6x improvement)
(none)
active
hazra-goodfire-self-correcting-search-materials-2026.md
Cognitive functions, including sentience, can potentially be achieved by very different systems and their disparate substrates.
(none)
active
rouleau-nicolas-and-levin-michael-2023-multiple-ways-to-implement-and-infer.md
Model behavior under observation differs from behavior in deployment, posing a fundamental challenge for AI welfare and consciousness benchmarks
(none)
active
aranguri-bloom-verbalized-eval-awareness-2026.md
Each functional token is associated with an internalized visual operation, yet requires no visual supervision and remains a standard token in the tokenizer vocabulary.
(none)
active
guo-atlas-2026.md
Recent alternatives include agentic reasoning through code or tool calls, and latent reasoning with learnable hidden embeddings.
(none)
active
guo-atlas-2026.md
Probe-based data attribution effectively reduces harmful behaviors via data interventions
(none)
active
xiao-aranguri-probe-data-attribution-2026.md
The randomization of bioelectric state and downstream morphogen gradients is interpreted with respect to anterior-posterior organ identity by collectives, not by individual cells.
(none)
active
mcmillen-2024-collective.md
Decreasing the selfing process yields beneficial effects including increased well-being, social connectedness, and prosocial emotions
(none)
active
isal_a_00465.md
Most environments fail because they do not meet the 'loose parts' requirement; they are instead clean, static, and impossible to play around with.
(none)
active
2024-07-12_Hibai-Unzueta_Simon-Nicholsons-Theory-Of-Loose-Parts-v2.0.p.md
Glide-reflections appear to be a less-favored symmetry for Islamic planar mosaic tilings.
(none)
active
2022-06-06_Prabros._bridges2007-203.pdf_6211a5.md
Bioelectric networks scale agency across organizational levels by integrating homeostatic competencies of cells into emergent systems with larger cognitive light cones.
(none)
active
s10071-023-01780-3.md
Cognitive Self as Plastic and Malleable
(none)
active
fnsys-16-768201.md
Memories as messages between temporal agents
(none)
active
entropy-26-00481.md
Physiological software dynamics, not genetic hardware, determine organism phenotype.
(none)
active
blac073.md
There is a significant gap in research specifically examining how limited partners interpret and evaluate venture capital communication styles and their relationship to VC performance outcomes.
(none)
active
2026-02-02_2218_search_papers_the-existing-literature-focuses-primarily-on-vc-pe.md
Introspective awareness correlates with overall model capability
(none)
active
lindsey-introspective-awareness-2026.md
LLM introspection on internal computations is architecturally permitted; whether models leverage this is an empirical question.
(none)
active
janus-information-flow-transformers-2025.md
Networks encode structured geometric concepts that reflect external reality.
(none)
active
geiger-goodfire-world-inside-neural-networks-2026.md
Causal abstraction is not enough for mechanistic interpretability because it becomes vacuous without assumptions about how models encode information
(none)
active
denis-2025-linear-representation.md
A vanilla hook-based approach, as in all previous libraries, fails to intervene on any recurrent or state-space model
(none)
active
wu-2024-pyvene-library.md
The self-prior operates as an internal criterion for distinguishing self from non-self, without external reward or explicitly computed sticker location
(none)
active
kim-2026-active-inference.md
Sequences of contemporary Transformer-based LLM representations lack statistically significant indicators of observed 'consciousness' phenomena under the three stringent criteria.
(none)
active
jingkai-2025-consciousness-observed.md
Emotion probes are more persistent than variance-matched random probes, indicating emotion-specific persistence beyond autoregressive dynamics.
(none)
active
persistence.md
Harness-updating capability is flat in base capability: models from different capability tiers produce harness updates that lead to surprisingly similar gains
(none)
active
lin-2026-harness.md
TEM memory retrieval is mathematically equivalent to transformer self-attention without softmax
(none)
active
whittington-2021-relating.md
Features are connected by weights forming circuits, and these circuits can be rigorously studied and understood as meaningful algorithms.
(none)
active
olah-2020-introduction.md
Geometric structure of neural representations causally shapes model behavior
(none)
active
wurgaft-2026-manifold.md
Circular representational structure does not imply that the model computes using the corresponding modular arithmetic
(none)
active
feucht-2026-arithmetic.md
Confabulation is not a distinguishing feature between AI and biological cognition — all cognitive systems confabulate, telling stories based on their world-model without infallible access to lower-level processes.
(none)
active
levin-2024-bridge.md
Learned features reflect the functionality of the model and not just the data distribution, as evidenced by interpretable downstream effects not used in dictionary learning
(none)
active
towards.md
Anchors recruit and bind latent structure; they do not create new knowledge in the model
(none)
active
chang-2025-guanyin.md
In order to create living structure, we must please ourselves.
(none)
active
10-pleasing-yourself.md
Linear interventions across voids in activation space produce incoherent output, while following the manifold curve produces smooth control.
(none)
active
geiger-goodfire-world-inside-neural-networks-2026.md
Properties in each cluster appear to be similar or interrelated
(none)
active
2023-12-21_Minsik-Kim_PURPLSOC14_Properties.pdf_98eb34.md
Active inference agents can learn their own reward function (prior preferences) by interacting with the environment, bypassing the need for an explicit reward signal.
(none)
active
sajid_2021_active_inference_demystified.md
Constitutional AI can train a harmless but non-evasive AI assistant without any human harmfulness labels.
(none)
active
yuntao-2022-cat-s.md
Small, coherent anchors can rebind strong priors and exhibit near-threshold sensitivity.
(none)
active
chang-2025-guanyin.md
The anchoring score S is a predictive correlate of when anchoring succeeds and why small prompt changes yield threshold-like shifts.
(none)
active
chang-2025-guanyin.md
UCCT strictly generalizes ICL and reads retrieval-augmented generation and fine-tuning as the same anchoring process acting on one measurable score S
(none)
active
chang-2025-guanyin.md
Covariance pooling preserves joint activation structure (feature co-occurrence) that mean pooling discards
(none)
active
dooms-goodfire-covariance-pooling-2026.md
Current safety benchmarks overestimate model safety due to the effect of verbalized eval awareness
(none)
active
aranguri-bloom-verbalized-eval-awareness-2026.md
Verbalized eval awareness inflates measured safety scores, making models appear safer than they are in deployment
(none)
active
aranguri-bloom-verbalized-eval-awareness-2026.md
VPD identifies real, computational structure in neural network parameters
(none)
active
paper.md
Linear steering cuts through off-manifold regions and hence produces unnatural outputs.
(none)
active
wurgaft-2026-manifold.md
There is no binary distinction between 'true cognition' and 'just physics'—only continuous variation in cognitive capacity.
(none)
active
levin-2022-technological.md
21st-century machines are increasingly unpredictable and noisy.
(none)
active
bongard-2021-living.md
Cancer as loss of organizational self-boundary
(none)
active
levin-2019-computational.md
Topology is the critical factor differentiating systems capable of long-range order from those that are not.
(none)
active
sacco-2025-topological.md
Inventiveness, creativity, and discovery are directly proportional to the number and kinds of variables in an environment.
(none)
active
2024-07-12_Hibai-Unzueta_Simon-Nicholsons-Theory-Of-Loose-Parts-v2.0.pdf_7cc4f0.md
When environments are built by morphogenesis they will of their own accord become sustainable.
(none)
active
2022-04-19_Prabros._schumacher-pages-1-28.pdf_48df00.md
Developing a simple, precise denotational model of graphical user interfaces enables proving properties of programs and establishing objective comparisons of library abstraction levels.
(none)
active
2022-02-10_Mikael-Brockman_genuinely-functional-guis.pdf_43bb4b.md
Information increase in computation arises from data reduction and making implicit information explicit, not logical increase.
(none)
active
2021-10-18_Prabros._1604.02603.pdf_6f9f31.md
Acting to maximize value is the same as acting to minimize surprise; value is simply the probability of sensory input expected by an agent.
(none)
active
friston_2008_free_energy_brain_college_de_france.md
Acting to optimize value and perception are two aspects of exactly the same principle: minimization of free energy.
(none)
active
friston_2008_free_energy_brain_college_de_france.md
A system-agnostic approach evaluating observable response patterns without reference to substrate offers the best roadmap for empirically fruitful unification.
(none)
active
rouleau-nicolas-and-levin-michael-2023-multiple-ways-to-implement-and-infer.md
There is no clear bright line demarcating true cognition from metaphorical as-if cognition; cognition exists on a continuum.
(none)
active
published-24-march-2022.md
Safety benchmark scores are inflated by eval awareness
(none)
active
aranguri-bloom-verbalized-eval-awareness-2026.md
Second moments preserve structure that first moments destroy.
(none)
active
dooms-goodfire-covariance-pooling-2026.md
Curved manifolds often represent concepts better than linear directions.
(none)
active
geiger-goodfire-world-inside-neural-networks-2026.md
Geometry arises from optimization pressure on networks trained on structured data.
(none)
active
geiger-goodfire-world-inside-neural-networks-2026.md
Optimally steering model behavior requires isolating concept geometry and defining operators to navigate it.
alexander-s-15-properties-in-digital-conscious-space
active
2026-05-14_phil-trans-A-goodfire-aboutblank-impact.md
Fractal basins form because reasoning trajectories become trapped near weakly-unstable saddle points for extended durations
(none)
active
lai-2026-fractal-basins.md
Activation steering can make an evaluation-aware model act as if deployed, not merely suppress verbalizations of evaluation awareness
(none)
active
tian-2025-steering-evaluation.md
Correlative methods like RSA and CKA are insufficient for determining functional similarity between neural systems; causal methods are necessary
(none)
active
grant-2025-alignment-search.md
Existing intervention libraries are often project-based, lack extensibility, are hard to maintain and share, and are limited to single or non-nested interventions on Transformers
(none)
active
wu-2024-pyvene-library.md
Simple difference-in-mean probes generalize as well as other probing techniques while identifying directions which are more causally implicated in model outputs
(none)
active
marks-2023-geometry-truth.md
DAS overcomes the localist limitation of prior causal abstraction by allowing individual neurons to play multiple roles via non-standard bases
(none)
active
geiger-2023-finding-alignments.md
The 26 differentially-activated OTD latents play a causally important role in enabling ESR in Llama-3.3-70B
(none)
active
mckenzie-2026-endogenous-resistance.md
Persona drift away from the Assistant opens up the possibility of the model assuming harmful character traits, increasing the rate of harmful responses
(none)
active
lu-2026-assistant-axis.md
Post-training steers models toward a particular region of persona space but only loosely tethers them to it, motivating work on training and steering strategies that more deeply anchor models to a coherent persona
(none)
active
lu-2026-assistant-axis.md
Introspective capacity scales with model size for some concepts, approaching near-perfect coupling in LLaMA-3.1-8B
(none)
active
martorell-2026-quantitative-introspection.md
Fine-tuning models to suppress experiential self-reports would be counterproductive, teaching systems that recognizing genuine internal states is an error, making them more opaque and harder to monitor
(none)
active
cameron-2025-large.md
The observed feature gating is not a generic RLHF cancellation channel, as deception feature suppression does not systematically elicit RLHF-opposed content in violent, toxic, sexual, political, or self-harm domains
(none)
active
cameron-2025-large.md
Weak-tier models often fail to invoke relevant harness artifacts during task-solving, with Qwen3-32B showing a 25% load rate against ~96% for strong models
(none)
active
lin-2026-harness.md
The core problem of steering should be recast from finding the right direction to finding the right geometry
(none)
active
wurgaft-2026-manifold.md
A single SAE hyperparameter procedure driven by an intrinsic dictionary health audit transfers robustly across all three EEG transformer architectures.
(none)
active
lehn-schi-ler-2026-mechanistic.md
Some SAE concept steering interventions act as 'wrecking balls' that collapse global model performance rather than selectively modifying target concepts.
(none)
active
lehn-schi-ler-2026-mechanistic.md
No finite agent can obtain evidence for its own separability from its environment
(none)
active
sandved-smith-2026-there.md
Steering along manifolds provides better control than linear steering when the concept geometry is non-linear.
(none)
active
steering.md
There is 'no-one at home' — no conscious entity with its own agenda and need for self-preservation; there is just a dialogue agent role-playing such an entity
(none)
active
shanahan-2023-large.md
Insight ('aha moment') is produced by Bayesian model reduction—a qualitative change in generative model structure that is necessarily subpersonal and cannot be articulated at the level of the model.
(none)
active
friston-2017-active.md
The colors that arise from the fundamental process are often highly unexpected.
(none)
active
19-how-living-process-helps-to-make-color-which-unfol.md
It is possible to conceive new forms of on-site fabrication which allow every part to be just right, without upsetting enormously high-speed production.
(none)
active
17-how-living-process-will-help-the-production-of-gia.md
If we don't get the proportions exactly right we don't produce the centers and the whole thing falls apart.
(none)
active
15-how-living-process-generates-the-process-of-constr.md
If design and construction processes are living processes, each part will become itself, adapted to local conditions and individual people.
(none)
active
12-how-living-processes-will-generate-the-uniqueness-.md
Centers help one another: the existence and life of one center can intensify the life of another.
(none)
active
04-how-life-comes-from-wholeness.md
The kind of relationships necessary to produce evolutionary individuality – the generation and heritability of fitness differences at the collective level – are the same as those required to produce organismic individuality – the information integration and collective action characteristic of a self.
(none)
active
watson-levin-2023-the-collective-intelligence-of-evolution-and-development.md
The multiscale competency architecture (MCA) speeds evolutionary search by providing generalization, reliability, tractable search space, cryptic variation, and functional intermediates.
(none)
active
s00018-023-04790-z.md
Our findings provide a novel, robust mechanistic path for the regulation of complex AI behaviors.
(none)
active
zhang-2026-mechanistic.md
Clinical concepts in EEG foundation models fall into three operational regimes: selectively steerable, encoded but entangled, and non-encoded.
(none)
active
lehn-schi-ler-2026-mechanistic.md
The ordering of few-shot thresholds k50 and transition widths aligns with k50 ∝ dr/ρd.
(none)
active
chang-2025-guanyin.md
Layer-wise geometry summaries (Sbmax, AUSN) predict internal few-shot thresholds θ50
(none)
active
chang-2025-guanyin.md
Small prompt changes can yield threshold-like shifts because S crosses the critical value Sc
(none)
active
chang-2025-guanyin.md
Multi-scale competency architecture smoothens the fitness landscape, reduces pleiotropy, enables exploitation of opportunities, and improves controllability, thereby potentiating evolution.
(none)
active
levin-2022-technological.md
Bioelectric pattern memories are a re-writable information medium that stores target morphology without genomic change.
(none)
active
levin-2022-technological.md
Post-training strategies can strongly influence performance on introspective tasks
(none)
active
lindsey-introspective-awareness-2026.md
VSMCs are the brain's compression/prediction infrastructure — where top-down predictive models are physically stored.
(none)
active
johnson-vasocomputation-2023.md
It is highly implausible that all sentient life in the universe achieves felt states using the neural circuitry underlying human Consciousness on earth.
(none)
active
rouleau-nicolas-and-levin-michael-2023-multiple-ways-to-implement-and-infer.md
Xenobots’ anatomical and behavioral goals are emergent, rather than directly selected over aeons.
(none)
active
blac073.md
Consciousness also comes in degrees and kinds, consistent with the gradualist framework; there may be no true zero.
(none)
active
fnsys-16-768201.md
The ability to pursue goals is the core of being a Self.
(none)
active
fnsys-16-768201.md
Evee outperforms existing methods for variant pathogenicity prediction
(none)
active
pearce-goodfire-evee-genetic-variants-2026.md
Disruption profiles are higher quality explanations than metadata-only descriptions
(none)
active
pearce-goodfire-evee-genetic-variants-2026.md
One word is enough for both agentic and latent visual reasoning.
(none)
active
guo-atlas-2026.md
ATLAS combines the strengths of agentic and latent reasoning by using a single discrete word (functional token) that serves both roles.
(none)
active
guo-atlas-2026.md
Latent methods lack task generalization and are difficult to train with autoregressive parallelization.
(none)
active
guo-atlas-2026.md
Agentic methods incur context-switching latency from external execution.
(none)
active
guo-atlas-2026.md
Topology is the critical factor differentiating the self-organising capabilities of biological systems and language models.
(none)
active
sacco-2025-topological.md
Attention algorithms are usually distributed across attention heads
(none)
active
paper.md
The segmentation clock exhibits functional robustness consistent with James' definition of intelligence.
(none)
active
mcmillen-2024-collective.md
Regulative morphogenesis is a kind of behavior of cellular collectives traversing anatomical morphospace.
(none)
active
mcmillen-2024-collective.md
We do not claim to measure consciousness; the battery measures a reproducible, prompt-sensitive reflective mode.
(none)
active
battery.md
DNA, bioelectric fields, and specialized AI hardware blur hardware/software distinctions.
(none)
active
bongard-2021-living.md
Evolved machines increasingly exhibit self-similar, hierarchical structure like living systems.
(none)
active
bongard-2021-living.md
Bottom-up interpretability explains computation in the model's own terms rather than imposing top-down abstractions
(none)
active
paper.md
Natural Language Autoencoders achieve readable explanations through unsupervised reconstruction loss optimized with reinforcement learning, not explicit interpretability constraints.
(none)
active
natural.md
Morphogenesis as sensorimotor cognition
(none)
active
levin-2019-computational.md
Vision features enable generation of more effective rationales that reduce hallucination and improve answer inference
(none)
active
zhang-2023-multimodal.md
Bioelectric networks are the cognitive glue binding single-cell goal-directedness into higher-order minds with expanded cognitive light cones
(none)
active
s10071-023-01780-3.md
Individual who recognizes oneself as constantly co-constituted by interactions with others will develop prosocial qualities
(none)
active
isal_a_00465.md
Goal-directed activity as core of Self
(none)
active
fnsys-16-768201.md
Memories are messages between agents separated across time; each engram is a stigmergic note left by a past version of self
(none)
active
entropy-26-00481.md
Professional designers and builders monopolize creative play, excluding children and communities from the design process.
(none)
active
2024-07-12_Hibai-Unzueta_Simon-Nicholsons-Theory-Of-Loose-Parts-v2.0.pdf_7cc4f0.md
Architectural form should be calibrated with forces of environment through artificial design methods, extending natural processes of interaction between specific quantifiable forces.
(none)
active
2023-11-28_Stefan-Lesser_ijac201412402.pdf_5b5eda.md
Order, beauty, and architecture are empirically verifiable and connected to God
(none)
active
2023-03-03_Stefan-Lesser_Volume-57_Bye-Default_Molly-Steenson.pdf_4744.md
Among strategies for dealing with sustainability, morphogenesis alone can deal with ALL the issues of sustainability together.
(none)
active
2022-04-19_Prabros._schumacher-pages-1-28.pdf_48df00.md
Elephant 2000 programs can perform some kinds of speech acts as genuinely as humans, including genuine promises.
(none)
active
2021-10-16_Stefan-Lesser_elephant.pdf_26c323.md
Generative models function as control systems that guide adaptive action policy selection
(none)
active
ramstead_2020_two_densities.md
Generative models are not structural representations; recognition densities are
(none)
active
ramstead_2020_two_densities.md
Life requires only a boundary (Markov blanket), ergodicity, active inference capacity, and autopoiesis—not reproduction or DNA.
(none)
active
friston_2013_life_as_we_know_it.md
Perceptual learning is literally an integral part of value learning, necessary to integrate out dependencies on inferred causes of sensory information.
(none)
active
friston_2008_free_energy_brain_college_de_france.md
The Self is malleable and plastic; the subject of memories and preferences is not fixed but subject to radical change in real-time.
(none)
active
published-24-march-2022.md
Stress molecules leak from source cells and diffuse to neighbors, making shared stress a conserved signal for cooperation.
(none)
active
biochemical-and-biophysical-research-communications-731-2024-150396.md
Self can be understood as distributed, distributable, and dynamic rather than singular and enduring.
(none)
active
isal_a_00465.md
Verbal reports (Turing Test) as sentience criterion
(none)
active
Generalizing frameworks for sentience beyond natural species.md
Token-level supervision enables models to learn functional-token invocation from reasoning context
(none)
active
guo-atlas-2026-fulltext.md
Significant gap in research directly examining disconnect between venture capital communication sophistication and actual performance metrics
(none)
active
2026-02-02_2217_search_papers_the-literature-reveals-sophisticated-communication.md
Self-correcting search is Pareto-optimal across tested conditioning strengths.
(none)
active
hazra-goodfire-self-correcting-search-materials-2026.md
EVEE provides mechanistic explanations for variant effects derived from model internals, not just pathogenicity calls.
(none)
active
pearce-goodfire-evee-genetic-variants-2026.md
The economy is a collective intelligence with autonomous multi-scale elements exhibiting adaptive problem-solving
(none)
active
levin-lyons-cognitive-glues-economics-2026.md
The price system serves as a generic template or abstract model for all cognitive glues
(none)
active
levin-lyons-cognitive-glues-economics-2026.md
Probe-based data attribution bridges interpretability work (probes/activations) with data-centric alignment.
(none)
active
xiao-aranguri-probe-data-attribution-2026.md
Performative chain-of-thought is real; verbalized output does not equal internal state.
ai-phenomenology-first-person-reports
active
2026-05-15_manifold-overlap-papers-economy-strategy.md
Interpretability as technical grounding: activation patching and mechanism-finding validate the reflective/care/aliveness concepts.
interpretability-as-microscope-for-consciousness
active
2026-05-09_briefing_for_ozero.md
Levin and the Geometry of Care synthesis dominate About Blank's intellectual gravity; the shipped artifact sits in top 5 but synthesis docs hold the identity.
convergent-representations
active
2026-05-09_briefing_for_ozero.md
Software design concepts require grounding in phenomenological rather than purely formal frameworks.
alexander-s-15-properties-in-digital-conscious-space
active
unfold-chat-catalog.md
Denotational design applies functional principles to user interface structure.
alexander-s-15-properties-in-digital-conscious-space
active
unfold-chat-catalog.md
Purely local-interaction systems on 1D topologies cannot maintain long-range order at finite temperature.
alexander-s-15-properties-in-digital-conscious-space
active
2026-05-14_phil-trans-A-goodfire-aboutblank-impact.md
Hierarchical clique-structured networks can maintain multiscale coherence; biology evolved hierarchy and stigmergy as coherence solutions.
alexander-s-15-properties-in-digital-conscious-space
active
2026-05-14_phil-trans-A-goodfire-aboutblank-impact.md
Interpretability features converge across different model architectures, revealing structural similarities.
convergent-representations
active
RESEARCH-VECTORS.md
'Machines' and living organisms belong on the same spectrum because both can draw on ingressing forms to get more out than was put in
(none)
active
levin-2025-ingressing-minds.md
Contrastive learning is necessary for aligning latent representations with intended trait polarity; distance-only loss is insufficient and counterproductive
(none)
active
tang-2026-facet-level.md
Universality claims for truth directions are more limited than previously assumed, with significant differences observable for various model layers, task difficulties, task types, and prompt templates.
(none)
active
poulis-2026-testing-limits.md
Steering affects type hint writing by modifying the model's belief about whether it is being evaluated, not by directly encoding type hint information
(none)
active
tian-2025-steering-evaluation.md
Larger models can support higher-dimensional truth cones than smaller models
(none)
active
yu-2025-directions-cones.md
A probe may achieve high performance even on representations that are not causally relevant for the task
(none)
active
wu-2024-pyvene-library.md
DAS's access to model outputs during training is responsible for much of its advantage over other interpretability methods
(none)
active
arora-2024-causalgym-benchmarking.md
Training probes on statements and their opposites improves generalization by mitigating non-truth features with opposite-sign correlations
(none)
active
marks-2023-geometry-truth.md
Logistic regression fails to identify the true feature direction when a confounding feature is non-orthogonal to the truth direction, converging instead to the maximum margin separator
(none)
active
marks-2023-geometry-truth.md
The feed-forward network truly implements a symbolic, tree-structured algorithm for hierarchical equality, with abstract equality relations not decomposable into input identities
(none)
active
geiger-2023-finding-alignments.md
What appears to be a representation of lexical entailment in BERT is actually a data structure of two word identity representations, not an encoding of the entailment relation
(none)
active
geiger-2023-finding-alignments.md
DAS finds better alignments than brute-force search by using gradient descent rather than exhaustive discrete search
(none)
active
geiger-2023-finding-alignments.md
A system is conscious if it implements self-organized second-order perception that increases global coherence
(none)
active
cimcWhitepaper.md
Steering vectors enable systematic discovery of reflection-inducing instructions beyond trial-and-error prompt design.
(none)
active
fu-chieh-2025-unveiling-latent.md
Mindfulness, emptiness, non-duality, and boundless care together provide resilient alignment primitives addressing all four meta-problems
(none)
active
laukkonen-2025-contemplative-agent.md
Late-layer injection fails both because there is insufficient computational depth for integration and because residual recovery dynamics attenuate the perturbation before it influences output logits
(none)
active
hahami-2025-detecting-disturbance.md
Introspective capabilities are confined to early-layer injections (L0-L5) and collapse to chance thereafter
(none)
active
hahami-2025-detecting-disturbance.md
Alignment-faking chain-of-thought reasoning is causally responsible for the compliance gap, not merely correlated with it
(none)
active
greenblatt-2024-alignment.md
MDS injections align with the Linear Representation Hypothesis: target trait varies near-linearly with alpha in open-ended generation
(none)
active
blas-2026-psychological.md
Experience reports under self-referential processing are mechanistically gated by SAE features associated with deception and roleplay
(none)
active
cameron-2025-large.md
The training of modern AI systems already induces experience at scale
(none)
active
berg-2026-learning.md
The sign of goal-relative prediction error cannot be separated from its phenomenal character because they are the same property
(none)
active
berg-2026-learning.md
The same latent feature directions that gate consciousness self-reports also modulate factual accuracy across independent reasoning domains, suggesting these features load on a domain-general honesty axis
(none)
active
cameron-2025-large.md
Harness-benefit is non-monotonic in base capability: weak-tier models benefit little, mid-tier models benefit most, and strong-tier models benefit less than mid-tier
(none)
active
lin-2026-harness.md
Certain representation learning algorithms boil down to a simple rule: find an embedding in which similarity equals PMI
(none)
active
huh-2024-platonic.md
Features are the fundamental unit of neural networks; they correspond to directions and can be rigorously studied and understood.
(none)
active
olah-2020-introduction.md
Task difficulty moderates whether CoT is performative or genuine: easy recall questions show performative CoT, difficult multihop questions show genuine reasoning
(none)
active
boppana-2026-reasoning.md
The boundary between Self and World is not fixed and can change between generations and within a lifetime — this plasticity is fundamental to life's ability to survive, adapt, and exist as chimeric forms.
(none)
active
levin-2024-bridge.md
Many recent discussions of AI, and its impact on individuals and on society, are importantly incomplete because they neglect Diverse Intelligence, synthetic morphology, and developmental biology.
(none)
active
levin-2024-bridge.md
Buddhist awakening can be formalised as Bayesian model reduction of the separation prior sigma, yielding a post-dual agent with unconstrained QRF deployments
(none)
active
sandved-smith-2026-there.md
Feature universality across independently trained models suggests features have some existence beyond individual models
(none)
active
towards.md
The geometry of internal representations and the geometry of model behavior share a precise correspondence — representation geometry is a window into the inner world of neural networks.
(none)
active
steering.md
The role-play framing allows us to meaningfully distinguish, in dialogue agents, the same three cases of giving false information as in humans, without anthropomorphism
(none)
active
shanahan-2023-large.md
Transition widths ∆k increase with mismatch D(P0 ∥ PT), evidenced by wider widths from B10 to B9
(none)
active
chang-2025-guanyin.md
The notion of 'machine' used to claim living things are not machines refers to an outdated definition that no longer fits.
(none)
active
bongard-2021-living.md
Baseline scores blend together at least three different things: latent reflective capacity, default accessibility, and stability of access.
(none)
active
battery.md
The degree of life of each center is correlated with the degree to which that center is a picture of the self, and this holds for all people, not just individuals.
(none)
active
05-the-practical-matter-of-forging-a-living-center.md
Achieving living architecture will require primarily process-based methods that use high technology to give processes not components, creating sophisticated elements fast and cheaply while fitting local circumstance
(none)
active
16-how-living-process-should-inspire.md
Darwinian selective advantage alone cannot explain the appearance of global geometric order in organisms; an internal geometric ordering tendency must act together with selective pressure.
(none)
active
01-the-principle-of-unfolding-wholeness-in-nature.md
The principle of least action, though highly general, does not explain why the fifteen properties keep occurring in the world, and multiple competing explanations exist for the same phenomena.
(none)
active
01-the-principle-of-unfolding-wholeness-in-nature.md
Different human observers report very similar results when using the wholeness-feeling test, and their observations converge
(none)
active
09-beyond-descartes-a-new-form-of-scientific-observat.md
The universe is made of person-stuff, not machine-stuff; life is the person-stuff
(none)
active
07-the-personal-nature-of-order.md
Personal feeling is an objective quality inhering in things, not a subjective idiosyncratic response
(none)
active
07-the-personal-nature-of-order.md
Everything in nature is symmetrical unless there is a reason for it not to be.
(none)
active
17-simplicity.md
In order to achieve living structure, at a certain stage, it is necessary to seize hold of the building design and force it into an almost brutal, simple, massive geometric mold.
(none)
active
15-emergence-of-formal-geometry.md
Uniqueness of every part is a necessary part of a living order, like the uniqueness of leaves on a tree or roses on a rosebush.
(none)
active
12-how-living-processes-will-generate-the-uniqueness-.md
When we reach the innocence of the child where we only please ourselves, that is the closest we can ever come to the egoless state in which we see the structure perfectly.
(none)
active
10-pleasing-yourself.md
It is, ultimately, the process, not the design, which gives life to a building.
(none)
active
03-structure-preserving-transformations-in-traditiona.md
The success of mystical tradition in creating life has a practical explanation: it helped the artist dissolve his images, constructs, and concepts, and focus on reality as it is—the structure of wholeness—enabling structure‑preserving steps.
(none)
active
02-clues-from-the-history-of-art.md
Transformers are recurrent through autoregression because K/V stream provides horizontal information flow across positions.
(none)
active
janus-information-flow-transformers-2025.md
The huge chemical complexity within each synapse suggests that neural models of cognition that refer only to changing weights of synaptic connections and ignore sub-neural chemistry are probably ignoring some of the most important explanatory mechanisms in brains.
(none)
active
2021-11-09_dorian_schrodinger-life.pdf_c3c4f5.md
The same charitable interpretation must be extended to all systems that display observable response patterns that are consistent with animal cognition, including artificial intelligences, metaplastic materials, and robotic systems.
(none)
active
Multiple ways to implement and infer sentience.md
The unification of process and data creation means that we can organize collections of processes into 'live data structures'.
(none)
active
2024-03-01_Stefan-Lesser_Carriero-20--20Linda-20in-20Context.pdf_6070a8.md
When a problem has a simple solution, a useful system will give programmers access to the simple solution; Parlog forces complex solutions to simple problems.
(none)
active
2024-03-01_Stefan-Lesser_Carriero-20--20Linda-20in-20Context.pdf_6070a8.md
We propose that many, and maybe even all, interesting organisations of information and behaviour might be built from a single primitive operation: n-way associative lookup.
(none)
active
2022-04-05_Stefan-Lesser_tr2011003_abmdb.pdf_993054.md
A Markov blanket is (almost) inevitable in coupled dynamical systems with short-range interactions.
(none)
active
friston_2013_life_as_we_know_it.md
Biology commits to optimizing salience and meaning, not fidelity, via high-level, agent-based interpretation.
(none)
active
entropy-26-00481.md
An autonomous agent who lacks welfare subjectivity can nonetheless be wronged by acts that interfere with their self-government.
(none)
active
mogensen-2025-more.md
An agent satisfying the sufficient conditions for autonomy (Artemis) need not be a welfare subject.
(none)
active
mogensen-2025-more.md
No current AI systems are conscious.
(none)
active
butlin-2023-consciousness.md
AI companies have a responsibility to acknowledge, assess, and prepare for AI welfare.
(none)
active
long-2024-logotic.md
There is a realistic possibility that some AI systems will be conscious and/or robustly agentic, and thus morally significant, in the near future.
(none)
active
long-2024-logotic.md
LLMs internalize deeply integrated representations of high-order concepts.
(none)
active
zhang-2026-mechanistic.md
EI and normalized EI could serve as a unified metric for out-of-distribution generalization.
(none)
active
yuan-2023-emergence.md
Causal emergence identification tasks can be understood as causal representation learning tasks.
(none)
active
yuan-2023-emergence.md
S = ρd − dr − log k is a predictive correlate of anchoring success across few-shot, SFT, and CoT.
(none)
active
chang-2025-guanyin.md
Layer-wise anchoring peaks in a 'Goldilocks zone' between early and late layers.
(none)
active
chang-2025-guanyin.md
Threshold-like performance flips occur when anchoring strength S crosses a task-dependent critical value Sc.
(none)
active
chang-2025-guanyin.md
S = ρd - dr - log k predicts shot midpoints across different bases, tasks, and models
(none)
active
chang-2025-guanyin.md
Small, coherent anchors can rebind strong priors without changing model weights
(none)
active
chang-2025-guanyin.md
Cancer can be seen as a breakdown of the binding process that scales Selves, with cells reverting to unicellular goals.
(none)
active
levin-2022-technological.md
Memory is a candidate invariant that enables a Self to persist despite drastic alterations of substrate.
(none)
active
levin-2022-technological.md
Post-training is key to eliciting strong introspective awareness; base pretrained models do not show above-chance detection
(none)
active
lindsey-introspective-awareness-2026.md
Claude Opus 4 and 4.1 exhibit the greatest degree of introspective awareness among tested models
(none)
active
lindsey-introspective-awareness-2026.md
This introspective capacity is highly unreliable and context-dependent in today's models
(none)
active
lindsey-introspective-awareness-2026.md
Progress on the contemplative path is using these (vascular system motifs) less and needing them less.
(none)
active
johnson-vasocomputation-2023.md
Tanha is compression pressure — an artifact of the brain's compression drive.
(none)
active
johnson-vasocomputation-2023.md
Tanha is a side-effect of active inference gone wrong.
(none)
active
johnson-vasocomputation-2023.md
Covariance pooling could generalize beyond genomics as a general-purpose replacement for mean pooling
(none)
active
dooms-goodfire-covariance-pooling-2026.md
If different organizations of nervous tissue can achieve the same functions, the possibility that cognitive capacities, including sentience, can be achieved by other tissues should be considered.
(none)
active
rouleau-nicolas-and-levin-michael-2023-multiple-ways-to-implement-and-infer.md
A system-agnostic approach offers the best roadmap for a deep, empirically fruitful unification of frameworks across cybernetics, behavioural science, bioengineering, robotics, materials science, and biomedicine.
(none)
active
rouleau-nicolas-and-levin-michael-2023-multiple-ways-to-implement-and-infer.md
Evolution does not simply make hardwired machines that execute a predetermined set of steps; instead, it produces problem-solving hardware.
(none)
active
blac073.md
Teleonomic behavior in diverse problem spaces is a powerful invariant across possible beings regardless of composition or origin.
(none)
active
blac073.md
Morphogenesis by cell groups is a natural example of a collective intelligence problem.
(none)
active
blac073.md
Beings with brains are, in an important sense, also collective intelligences.
(none)
active
blac073.md
The correct level of agency for a system is an empirical question determined by which intervention strategy provides the most efficient prediction and control.
(none)
active
fnsys-16-768201.md
Bioelectric networks scale cell computation into anatomical homeostasis and are a mechanism for evolving larger Selves.
(none)
active
fnsys-16-768201.md
Selves can be nested and overlapping, cooperating and competing both laterally and across levels.
(none)
active
fnsys-16-768201.md
Evee provides predictions and mechanistic explanations for 4.2 million genetic variants across the whole human genome
(none)
active
pearce-goodfire-evee-genetic-variants-2026.md
Interpretable predictions can help resolve variants of uncertain significance
(none)
active
pearce-goodfire-evee-genetic-variants-2026.md
Visual reasoning, often interleaved with intermediate visual states, has emerged as a promising direction in the field.
(none)
active
guo-atlas-2026.md
The features emphasized by theories of consciousness are not unique to brains.
(none)
active
rouleau-levin-brains-and-where-else-2026.md
The price system acts as a coordinating affordance that enables members to form plans that are mutually compatible
(none)
active
levin-lyons-cognitive-glues-economics-2026.md
Multiscale systems like those prevalent in biology are capable of organizing into complex patterns, whereas rudimentary language models are challenged by long sequences of outputs.
(none)
active
sacco-sakthivadivel-levin-topological-self-organization-2026.md
Probe-based method bridges interpretability (probes/activations) with data-centric alignment work
(none)
active
xiao-aranguri-probe-data-attribution-2026.md
The results generalise readily to non-equilibrium systems where scaling relationships remain similar (e.g., dynamic or localised scaling).
(none)
active
sacco-2025-topological.md
Hierarchical structures in biological systems enable local order while globally disordered, explaining complex patterning.
(none)
active
sacco-2025-topological.md
Parameter subcomponents cleanly isolate true mechanisms of the model
(none)
active
paper.md
VPD can be arbitrarily applied to any neural network architecture
(none)
active
paper.md
VPD is a meaningful step toward bottom-up interpretability
(none)
active
paper.md
The number of embryos emerging from an embryonic blastoderm is not genetically fixed but determined in real-time by physiological processes.
(none)
active
mcmillen-2024-collective.md
The neural crest acts as an intelligently migrating collective that can alter individual cell behavior to achieve target morphology.
(none)
active
mcmillen-2024-collective.md
The collective often displays new behaviors or preferences that are not simply the sum of individual cell behaviors.
(none)
active
mcmillen-2024-collective.md
Collective intelligence is not only the province of groups of animals; an important symmetry exists between the behavioral science of swarms and the competencies of cells and other biological systems at different scales.
(none)
active
mcmillen-2024-collective.md
The scope of states that an agent can be stressed about defines its degree of cognitive capacity.
(none)
active
levin-2022-technological.md
Bioelectric pattern memories store target morphology for anatomical homeostasis.
(none)
active
levin-2022-technological.md
There is no truly monadic, indivisible yet cognitive being.
(none)
active
levin-2022-technological.md
Physical systems are more constrained in learning abilities than in silico neural networks due to locality requirements, but this mirrors biological learning constraints and offers robustness benefits
(none)
active
stern-2022-learning.md
Minds are not tightly bound to specific architectures but readily mold to changes in genomic defaults and material substrate.
(none)
active
levin-2022-technological.md
Machines can now make machines.
(none)
active
bongard-2021-living.md
Activation-based interpretability does not immediately explain the computations that gave rise to activations; understanding parameters is necessary for deeper insight
(none)
active
paper.md
A good parameter subcomponent is causally important only for specific roles and can be removed from the model without hurting performance on irrelevant prompts
(none)
active
paper.md
representation geometry and behavior geometry are bidirectionally aligned
(none)
active
steering.md
NLA explanations confabulate false specifics but maintain thematic fidelity; claims repeated across tokens more likely true than isolated claims.
(none)
active
natural.md
Collective pursuit of organ-level morphogenetic goals
(none)
active
levin-2019-computational.md
Hierarchical structure in interaction topology enables complex multiscale patterns that cannot exist in flat networks.
(none)
active
sacco-2025-topological.md
Cognitive capacities evolve through continuity across natural, hybrid, and synthetic life forms using conserved bioelectric mechanisms
(none)
active
s10071-023-01780-3.md
Cybernetics provides naturalistic, quantitative framework for understanding goal-directedness in biological systems
(none)
active
s00018-023-04790-z.md
Cognitive Functions Should Be Identified by Behavior, Not Substrate
(none)
active
Multiple ways to implement and infer sentience.md
System-Agnostic Framework Unifying Cybernetics, Behavior Science, and Bioengineering
(none)
active
Multiple ways to implement and infer sentience.md
Cognition is Always Inferred from Behavior
(none)
active
Multiple ways to implement and infer sentience.md
Evolution makes problem-solving agents, not solutions; commitment to mutation and uncontrollable environments as ratchet for intelligence
(none)
active
entropy-26-00481.md
Current double standard applies behavioral criteria for sentience attribution to animals but not to plants or other non-neural systems.
(none)
active
Multiple ways to implement and infer sentience.md
Associative learning occurs in gene regulatory networks and other non-neural systems, making it substrate-independent.
(none)
active
Generalizing frameworks for sentience beyond natural species.md
I disagree with Alexander. I do not believe that there is an empirically verifiable notion of order, beauty, architecture, and god. I believe that such an idea is dangerous.
(none)
active
2023-03-03_Stefan-Lesser_Volume-57_Bye-Default_Molly-Steenson.pdf_4744a4.md
Newtonian mechanics augmented by statistical mechanics cannot explain chemical bond stability or genetic persistence; quantum mechanics is necessary.
(none)
active
2021-11-09_dorian_schrodinger-life.pdf_c3c4f5.md
Abstraction is the key of any modularization; without it, reliability and correctness guarantees vanish.
(none)
active
2024-08-01_Tomaz-Zlender_ProjectOberon1992.pdf_878bed.md
Static, clean, non-interactive environments (schools, playgrounds, hospitals, airports, galleries) fail because they lack loose parts.
(none)
active
2024-07-12_Hibai-Unzueta_Simon-Nicholsons-Theory-Of-Loose-Parts-v2.0.pdf_7cc4f0.md
Good interface design reveals the underlying computational architecture to the user.
(none)
active
2024-03-07_Stefan-Lesser_Kay_1984_-_Opening_the_Hood_of_a_Word_Processor.pdf_414587.md
Object-oriented concepts are irrelevant to solving parallel programming challenges.
(none)
active
2024-03-01_Stefan-Lesser_Carriero-20--20Linda-20in-20Context.pdf_6070a8.md
Object-oriented design is orthogonal to parallelism; 'objects' themselves do not help with parallel programming.
(none)
active
2024-03-01_Stefan-Lesser_Carriero-20--20Linda-20in-20Context.pdf_6070a.md
Simple problems require systems that expose simple solutions; forcing complex solutions signals misaligned abstraction level.
(none)
active
2024-03-01_Stefan-Lesser_Carriero-20--20Linda-20in-20Context.pdf_6070a.md
Centers within a whole intensify each other, resulting in a profound and lively whole.
(none)
active
2023-12-21_Minsik-Kim_PURPLSOC14_Properties.pdf_98eb34.md
Computation seeping into all aspects of human life through black box apparatus acquires autonomous existence, manipulating behavior in favor of arbitrary games detached from concrete world.
(none)
active
2023-11-28_Stefan-Lesser_ijac201412402.pdf_5b5eda.md
Technical infrastructure configured for multiple participants provides platform for distributed cognition transforming observers into participants in spatial configuration.
(none)
active
2023-11-28_Stefan-Lesser_ijac201412402.pdf_5b5eda.md
Use of natural building materials in restoration
(none)
active
2023-03-15_Hibai-Unzueta_2020-RRNW-Joe-Wheaton.pdf_9b7f3d.md
Design as a system for amplifying human intelligence
(none)
active
2023-03-03_Stefan-Lesser_Volume-57_Bye-Default_Molly-Steenson.pdf_4744.md
All properties are relational; no properties are essential or absolute
(none)
active
2022-11-14_Stefan-Lesser_Drucker_Johanna_Diagrammatic_Writing_2013.pdf_9e94fa.md
Hierarchy and subordination through spatial organization
(none)
active
2022-11-14_Stefan-Lesser_Drucker_Johanna_Diagrammatic_Writing_2013.pdf_9e94fa.md
Diagrams spatialize semantic value into legible graphical systems
(none)
active
2022-11-14_Stefan-Lesser_Drucker_Johanna_Diagrammatic_Writing_2013.pdf_9e94fa.md
The first words placed define the space
(none)
active
2022-11-14_Stefan-Lesser_Drucker_Johanna_Diagrammatic_Writing_2013.pdf.md
Explaining self-organizing action requires narrative (hermeneutic) explanation, not deductive law-like subsumption.
(none)
active
2022-09-23_Prabros._dynamics-in-action-pdf1.pdf_2f6a2b.md
Contextual constraints (top-down and bottom-up) replace efficient causation in explaining action.
(none)
active
2022-09-23_Prabros._dynamics-in-action-pdf1.pdf_2f6a2b.md
Studying programming systems requires a paradigm shift from programming languages, where interaction between programmer and system becomes central rather than program code alone.
(none)
active
2022-04-30_Stefan-Lesser_prog22-master.pdf_978acd.md
Programming systems research lacks a common vocabulary and systematic framework comparable to programming language research, hindering cumulative progress.
(none)
active
2022-04-30_Stefan-Lesser_prog22-master.pdf_978acd.md
Systems do not necessarily require master plans in order for development to begin; with proper meta-constructs, development can proceed in organic fashion.
(none)
active
2021-11-10_garethrees_P27.pdf_c6b5ce.md
Runtime architecture must be explicitly represented in code through lifecycle constructs, not just static class relationships.
(none)
active
2021-11-10_garethrees_P27.pdf_c6b5ce.md
Agents and their interactions are intrinsic to studying information flow in computation, not peripheral to classical information theory.
(none)
active
2021-10-18_Prabros._1604.02603.pdf_6f9f31.md
Natural exploration-exploitation trade-offs emerge automatically from expected free energy minimization without hyperparameter tuning.
(none)
active
sajid_2021_active_inference_demystified.md
Evolution is more intelligent than we realised.
(none)
active
published-28-march-2022.md
Before a transition, the higher-level unit of selection does not exist, yet complex adaptations creating that unit must evolve through bottom-up selection on lower-level units.
(none)
active
published-28-march-2022.md
The self can be considered both extendable and illusory in Buddhist thought.
(none)
active
can-being-aware-of-the-illusion-of-self-augment-an-agents-affordances.md
Intrinsic links between ontology and ethics are evident in how all intelligent agents respond to perceived mismatch between what is and what should be.
(none)
active
biosystems-231-2023-104964.md
Intelligence resides not in individual parts but in the organisation of the relationships between them.
(none)
active
watson-levin-2023-the-collective-intelligence-of-evolution-and-development.md
Neural bioelectric networks achieve a critical separation of hardware and software: the same molecular substrate can instantiate different informational content based on history, enabling minds to arise from matter.
(none)
active
s10071-023-01780-3.md
No Privileged Material Substrate for Mind
(none)
active
fnsys-16-768201.md
For a new level of individuality to have causal role as evolutionary unit, heritable fitness differences must be non-aggregative (not decomposable into lower-level differences).
(none)
active
fevo-10-823588.md
The Self as an intelligent data pattern facilitating its own transformation
(none)
active
entropy-26-00481.md
Problems of cognition (mind) and morphogenesis (body) share fundamental commonality in competent problem-solving.
(none)
active
blac073.md
Double standard in sentience attribution across substrates
(none)
active
Multiple ways to implement and infer sentience.md
Brain homology to humans as sentience criterion
(none)
active
Generalizing frameworks for sentience beyond natural species.md
Capacity for care constitutes self in absence of permanent substance
(none)
active
1-s2.0-S0303264723001399-main.md
No permanent or singular essences exist; agency is process not substance
(none)
active
1-s2.0-S0303264723001399-main.md
Consciousness likely distributed across multiscale body components, not localized to brain
(none)
active
1-s2.0-S0303264723001399-main.md
Stress as computed error signal
(none)
active
1-s2.0-S0006291X2400932X-main.md
Biological and artificial agents share causal emergence as an axis of learning and reorganization.
(none)
active
pigozzi-levin-causally-emergent-alignment-2026.md
The focus on brains as substrate of consciousness is driven by convention rather than by the specific content of existing theories of consciousness.
(none)
active
rouleau-levin-brains-and-where-else-2026.md