hypothesis
active
hypothesis:a-chronically-represented-character-emotional-state-if-it-exists-is-nonlinear-or-stored-in-key-value-cacheA chronically represented character emotional state, if it exists, is nonlinear or stored in key-value cache
Explains negative probing results.
Source paper
extracted_from(2026) · Nicholas Sofroniew · Isaac Kauvar · William Saunders · Runjin Chen +12
Neighborhood — ranked by edge-count
Papers (1)
paper
Findings (1)
finding
- Mixed LR probe achieves 0.386 to 0.826 accuracy against 6.7% chance, but fails on natural documentssupportsNo evidence of chronic emotion representation.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Interpretive hypothesis offered to explain why emotion features are more persistent
- Question raised by Anthropic and partially addressed by this paper's persistence evidence
- We hypothesize that persistently active emotional state representations exist in LLMs but may be missed by standard probing methods.hypothesis0.772Open hypothesis from the Anthropic paper that motivates this work
- Proposed explanation for why emotion probes are more persistent than variance-matched random probes
- Proposed mechanistic explanation for why emotion features are more persistent
- Central interpretive claim of the paper supported by multiple convergent analyses
- Main conclusion about the temporal dynamics of emotion features
- Interpretation of Part 2.