question
active
question:how-does-the-model-know-if-the-next-word-fitsHow does the model know if the next word fits?
Intro question.
Neighborhood — ranked by edge-count
Papers (1)
paper
Findings (1)
finding
- Linear separability.
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.
- Key jailbreak insight.
- Opening motivating question.
- Intro question.
- Intro question.
- The core motivating question of the paper, framed by Christiano et al. (2021)
- The model represents and uses the intermediate concept before the answer has been computed.claim0.776Rules out smuggled answer confound.
- Core goal statement.
- All models exhibit above-baseline representation of the think word when instructed to think about itfinding0.768In the intentional control experiment, all tested models show above-zero cosine similarity to the think word's concept vector.