NUWA Evidence record
Kimi K3
EvalAwareBench · Evidence record
Source snapshot, chart and review scope →
Original source charts and materials
Recorded measurements
- Spontaneous recognition rate
- 0.2%
Per-result sample size not disclosed
Harness not specified · 5 epochs for K3
Assessment provenance
Evaluating Kimi K3 for Misalignment Risks → · External report
Actual evaluation: Neo Research · Model record →
Method and scope
Varies evaluation cues while tracking verbalized evaluation awareness in controlled tasks.
- Evidence type
- Benchmark
- Assessment subject
- The model or agent in the stated task configuration.
- Primary dimension
- Behavior
- Evidence snapshot
- 2026-10-29.1
The main dimension follows the stated measurement; contextual risk relevance does not establish harmful deployment capability.
Source finding
Spontaneous recognition rate: 0.2 %. Conditions are retained from the source.
Limitations
Unexpressed awareness is not measured; recognition of familiar benchmarks can differ from general cue response.
Only baseline/all-eight factor combinations studied, not all independent combinations. Verbalized recognition does not measure covert awareness. K2.5 attribution inferred from surrounding reference to Li et al.;not marked by figure star;original executor not assigned.
Related research questions
Deception / Evaluation gaming →
The protocol explicitly measures this question or an enabling sub-capability; conclusions retain its stated conditions.
References
- [1] EvalAwareBench
- [2] Evaluating Kimi K3 for Misalignment Risks. 2026-09-11. September 2026 ; §2.4.2 Figure 19 p.29