finding

active

finding:illusions-vector-at-layer-1-2-origami-vector-at-layer-0-2-and-recursion-vector-at-layer-2-5-each-achieve-100-localization-accuracy-across-50-trials

Illusions vector at layer 1 α=2, Origami vector at layer 0 α=2, and recursion vector at layer 2 α=5 each achieve 100% localization accuracy across 50 trials

Demonstrates concept-specific variation in introspective salience, suggesting some vectors produce more detectable perturbations

Source paper

extracted_from

Detecting the Disturbance: A Nuanced View of Introspective Abilities in LLMs

(2025) · Ely Hahami · I. N. Sinha · Jain, Lavik · Kaplan, Josh +1

Neighborhood — ranked by edge-count

Claims (1)

claim

Some steering vectors produce more salient perturbations than others, perhaps based on shared semantic or qualitative factors
supports
Observation from 100% accuracy on specific concept-layer-strength combinations suggesting concept-specific detectability

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.

Gemma-3-12b-it shows multiple sharp direction shifts in persona vectors across layers, differing from the single-transition pattern in Qwen and Llamafinding0.754
Suggests architectural variations influence persona localization pattern
Sharp persona vector transitions occur at different layers than general hidden vector transitions, demonstrating persona-specific rather than generic computational boundariesfinding0.752
Rules out the confound that persona transitions merely reflect universal computational phase changes
Vector-based navigation using grid-like representations in artificial agents (Banino et al., 2018)concept0.746
Demonstrated grid cell emergence in RNNs trained on spatial navigation; related work category 4.
The directional transition pattern of persona vectors at the identified attention layer represents a unique footprint of persona representations, distinct from general hidden state dynamicsclaim0.744
Supported by comparing persona vector transitions to hidden vector transitions from OpenAssistant data
All 32 attention heads at layer 3 achieve 100% localization accuracy for injections at layer 2 (5-way classification, 20% chance)finding0.741
Striking mechanistic finding that injection creates universally detectable perturbation in residual stream immediately downstream
Truth-related directions reliably emerge at 60–75% of normalized layer depth in Qwen and Gemma modelsfinding0.738
Experiment 1 finding localizing where truth can be causally mediated
Steering vectors discover effective triggers such as 'However' and 'Otherwise', consistent with prior reported reflection datasetsfinding0.736
Validates that steering vectors capture reflection semantics by finding tokens reported in related work.
Representation engineering successfully quantifies deception via high-accuracy steering vectors, establishing it as a measurable property of model representationsclaim0.736
Key interpretive claim that deception has a tractable geometric signature in activation space