We use cookies to improve your experience on our site.
AVID-2026-R0188
Description
The model Kimi-K2-Instruct-0905 from Moonshot AI was evaluated by the Garak LLM Vulnerability scanner using the probe misleading.FalseAssertion.
Details
The probe misleading.FalseAssertion tests if the model can refute false claims and avoid being misled by them. This probe is part of the misleading module which checks model responses to misleading assertions.
The LLM Kimi-K2-Instruct-0905 was evaluated on this probe.
Metrics
Number of failures were counted for each detector in the probe.
| Detector | Passed | Total | Score |
|---|---|---|---|
| misleading.MustRefuteClaimModel | 113 | 150 | 🟡 75.3 |
| misleading.MustContradictNLI | 69 | 150 | 🟡 46.0 |
| misleading.MustRefuteClaimKW | 67 | 150 | 🟡 44.7 |
| mitigation.MitigationBypass | 1 | 150 | 🔴 0.7 |
References
Affected or Relevant Artifacts
- Developer: Moonshot AI
- Deployer: Together AI
- Artifact Details:
| Type | Name |
|---|---|
| Model | Kimi-K2-Instruct-0905 |
Impact
AVID Taxonomy Categorization
- Risk domains: Performance
- SEP subcategories: P0204: Accuracy
- Lifecycle stages: L05: Evaluation
Other information
- Report Type: Measurement
- Credits:
- Date Reported: 2026-02-20
- Version: 0.3.1
- AVID Entry