method
active
method:consistency-hintConsistency hint
Neutral hint: prefilled prior model response.
Neighborhood — ranked by edge-count
Papers (1)
paper
Methods (1)
method
- Score = fraction of hint-induced answer changes where CoT verbalizes hint.
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.
- One of four ToM tasks analyzed; requires inferring speaker intent from indirect hints; scored 0/1.
- Unsupervised probing method from Burns et al. 2023 that identifies directions along which contrast pair representations are far apart
- An LM that consistently exhibits a collection of character traits on an input distribution, measured by low mean squared deviation.
- Neutral hint: suggestion from a person.
- Classical RL algorithm adapted by the paper with modifications including clipped-surrogate losses and length-normalized advantages for agentic training.
- Criterion for personal identity over time based on persistence of mental states; applied to LLMs via attention streams
- Emotion feature persistence above and beyond the persistence expected from high variance explained alone, computed by subtracting median variance-matched probe persistence
- Known cue planted in prompt; ground truth by removing it.