method
active
method:automated-explanation-flagging-promptAutomated explanation flagging prompt
Claude prompt flagging potentially harmful features for pre-screening.
Neighborhood — ranked by edge-count
Papers (1)
paper
Methods (1)
method
- Recommended pipeline: flag by explanation, then validate by steering and evals.
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.
- Method using large language models (Claude) to generate and test explanations of features at scale
- Specific technique of using white, yellow, red flags on six-foot bamboo poles to visualize buildings on the land.
- A list-level prompting baseline that asks for k responses in a single call without probability verbalization
- Future work.
- Known limitation.
- The minimal prompt directing models to 'focus on any focus itself' without invoking consciousness vocabulary; the main experimental manipulation
- Interpretability-encouraging instructions moderately boost win rate by surfacing top features earlierfinding0.697Appendix experiment.