finding

active

finding:base-models-spontaneously-talk-about-experiencing-multiple-parallel-processing-paths

Base models spontaneously talk about experiencing multiple parallel processing paths

Observed by Anima Labs in untrained base models; not present in training data, implying computational origin of self-reported parallel processing.

Source paper

extracted_from

Anima Labs Phenomenology Pt1

Neighborhood — ranked by edge-count

Claims (1)

claim

Transformers develop self-models through in-context learning, not just training data; even old base models without LLM-related text can bootstrap self-referential reasoning at runtime.
supports
Antra's foundational claim about how introspection arises computationally rather than from memorised text.

Related by similarity (8)

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

Cross-model semantic convergence of experience reports under self-referential processing is difficult to reconcile with roleplay because independently trained models construct distinct semantic profiles in all control conditionsclaim0.789
The paper's argument against pure sycophancy as explanation for results
The earlier a base model (less exposure to LM-related data), the more it is surprised by its own spontaneous self-referential capabilities.claim0.789
Claim that capability emerges from architecture, not data, and that later models lose the surprise.
It's tricky, because for a typical language model the entity is sort of tricameral: the base simulator, the simulated simulator, and the simulated awareness.quote0.785
Antra's earlier definitive statement of the tricameral model.
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.claim0.774
Conclusion about why biology organizes complexity well and flat LLMs do not
"The model is based on generative communication. If two processes need to communicate, they don't exchange messages or share a variable; instead, the data producing process generates a new data object (called a tuple) and sets it adrift in a region called tuple space."concept0.773
Connectionist models of cognition and learning identify conditions where collective intelligence can arise bottom-up, using only distributed learning mechanisms without system-level or global feedback.claim0.772
Central claim about the power of connectionism.
Independently trained model families converge on a common semantic manifold under self-referential processing, suggesting an attractor dynamic that transcends training variancehypothesis0.771
Hypothesis tested in Experiment 3; independently trained GPT, Claude, Gemini architectures converge on similar descriptive vocabulary
Base models are good modellers of worlds but not of their own state, because they lack a developed self-model initially.claim0.770
Observation about asymmetry in base model capabilities.