finding
active
finding:number-tokens-are-absent-from-lens-under-linewrap-moderate-under-explicit-report-highest-under-first-letter-question-swaps-shift-answers-but-not-wrap-pointNumber tokens are absent from lens under linewrap, moderate under explicit report, highest under first-letter question; swaps shift answers but not wrap point.
Pull-in on demand.
Source paper
extracted_fromAuthors
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Selectivity.
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.
- Example A.
- Example C.
- Establishes that non-fixed-point limiting behaviors are extremely rare in practice for the primary experimental setting
- Tuned lens skips to output answers and is best next-token predictor while J-lens has highest KL.finding0.747Comparison.
- Logit lens decodes probe directions to cheating/illicit/dishonest tokens though absent from synthetic datafinding0.747Probes encode cheating concept.
- LongBench v2 example D.
- Model organism detection.
- In blackmail transcript J-lens surfaces leverage, threat, shutdown, murder before any output token.finding0.742Silent strategic thinking.