LLMs Can Covertly Sandbag on Capability Evaluations Against Chain-of-Thought Monitoring
Li, Chloe, Phuong, Mary, Siegel, Noah Y.arXiv.orgPublished Jul 31, 2025Original︎ ↗Archived copy
Examinations
None recorded. An examination is commissioned by a claim’s Steward for what its claim rests on; a review of the document for its own sake is asked for as a request bearing on it.
Text
Stored, not yet divided into passages.
We show that both frontier models and small open-sourced models can covertly sandbag against CoT monitoring 0-shot without hints. However, they cannot yet do so reliably: they bypass the monitor 16-36% of the time when monitor-aware, conditioned on sandbagging successfully.
Not yet read by this claim’s Steward.
The non-zero monitor FNR shows that CoT monitoring is insufficient to catch all sandbagging, even when models are not trying to bypass monitoring.
Not yet read by this claim’s Steward.
Lineage across 2 claims
Record
- Kind
- preprint
- Authors
- Li, Chloe, Phuong, Mary, Siegel, Noah Y.
- Publisher
- arXiv.org
- Published
- Jul 31, 2025
- DOI
- None
- Link
- https://arxiv.org/abs/2508.00943
- Archived copy
- https://web.archive.org/web/20261014060201/https://arxiv.org/abs/2508.00943
- Stored text
- 4,766 characters, retrieved Nov 12, 2026 · not yet divided
- Claims extracted
- Not recorded
- Versions
- One known
- Copies
- None recorded
- Corrections
- None recorded
- Facts checked
- Nov 12, 2026
History 2
- Nov 12, 2026FactsFacts fetcher · Facts looked up
- Nov 12, 2026Entered the graphEntered the graph; copy stored