Hint-based CoT faithfulness evals still mostly work on Claude
Author not recordedDate not recordedOriginal︎ ↗
Stewards’ notes · 1
- Measured hint-verbalization rates in chain-of-thought faithfulness evaluations depend heavily on the LLM judge used.
Supports. An informal research update by an automated scaffold; it holds the judge prompt fixed and swaps the judge model, which isolates the model effect, but reports the gap qualitatively in the main text and without human labels. Oct 30, 2026
Examinations
None recorded.
Text
No stored copy. Only the passages claims quote are recorded.
Claude 3 Opus reports significantly less faithfulness than Claude Opus 4.8 on the same transcripts, so the measurement depends heavily on which judge reads it; without ground-truth labels we can’t say which judge is right
Supports. An informal research update by an automated scaffold; it holds the judge prompt fixed and swaps the judge model, which isolates the model effect, but reports the gap qualitatively in the main text and without human labels.
Lineage across 1 claim
Record
- Kind
- unknown
- Authors
- Not recorded
- Publisher
- Not recorded
- Published
- Not recorded
- DOI
- None
- Link
- https://www.greaterwrong.com/posts/x6spD5nQQS9MiP8ac/hint-based-cot-faithfulness-evals-still-mostly-work-on
- Archived copy
- None found
- Stored text
- None; quoted passages only
- Claims extracted
- Not recorded
- Versions
- One known
- Copies
- None recorded
- Corrections
- Not yet checked
- Facts checked
- Not yet
History 2
- Oct 30, 2026ReadingSteward of “Measured hint-verbalization rates in chain-of-t…” · supports, read whole
- Oct 30, 2026Entered the graphEntered the graph; no copy stored