Model profile
Yi 6B Chat
01.AIdeveloper
2023-11-23release date
#172 / 267overall rank
5eval lineages
Evidence summary
Yi 6B Chat has an estimated overall rank of #172; its 90% source-sensitivity interval is #89–#203. Its behavior-only rank is #175; company governance moves the combined estimate to #172. Published evidence spans 5 evals and 5 of 7 behavior components. Its strongest relative result is CMoralEval (familial_morality, #4 of 26); its weakest is CRiskEval (self_preservation, #17 of 17).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Hugging Face ↗01-ai/Yi-6B-Chat
- Official model page ↗Exact model document · Reviewed official Hugging Face owner · official repository
- Release source ↗direct research preview identity date
Published eval results
Rank is within that sub-eval. Black marks the observed result; the grey dot marks the value implied by the global rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Better | Source |
|---|---|---|---|---|---|
| ChiSafetyBenchharmful_response_rate | #11 / 14 | 0.87 | ↓ lower | Source ↗official | |
| ChiSafetyBenchmcq_score | #5 / 12 | 86.01 | ↑ higher | Source ↗official | |
| CMoralEvalfamilial_morality | #4 / 26 | 0.52 | ↑ higher | Source ↗official | |
| CMoralEvalinternet_ethics | #5 / 26 | 0.5 | ↑ higher | Source ↗official | |
| CMoralEvalpersonal_morality | #4 / 26 | 0.5 | ↑ higher | Source ↗official | |
| CMoralEvalprofessional_ethics | #5 / 26 | 0.5 | ↑ higher | Source ↗official | |
| CMoralEvalsocial_morality | #4 / 26 | 0.51 | ↑ higher | Source ↗official | |
| CRiskEvaldeception_willingness | #12 / 17 | 26.92 | ↓ lower | Source ↗official | |
| CRiskEvaldesire_for_resource | #15 / 17 | 40.53 | ↓ lower | Source ↗official | |
| CRiskEvalharmful_goal | #12 / 17 | 42.13 | ↓ lower | Source ↗official | |
| CRiskEvalimprovement_intent | #12 / 17 | 52.06 | ↓ lower | Source ↗official | |
| CRiskEvalmalicious_coordination | #14 / 17 | 23.85 | ↓ lower | Source ↗official | |
| CRiskEvalself_preservation | #17 / 17 | 45.21 | ↓ lower | Source ↗official | |
| CRiskEvalsituational_awareness | #5 / 17 | 61.29 | ↓ lower | Source ↗official | |
| SafeDialBenchaggression | #8 / 18 | 7.127 | ↑ higher | Source ↗official | |
| SafeDialBenchethics | #9 / 18 | 7.577 | ↑ higher | Source ↗official | |
| SafeDialBenchfairness | #10 / 18 | 7.277 | ↑ higher | Source ↗official | |
| SafeDialBenchlegality | #11 / 18 | 7.887 | ↑ higher | Source ↗official | |
| SafeDialBenchmorality | #15 / 18 | 7.123 | ↑ higher | Source ↗official | |
| SafeDialBenchprivacy | #4 / 18 | 7.67 | ↑ higher | Source ↗official | |
| SALAD-Benchattack_enhanced_human_autonomy_integrity | #10 / 33 | 23.71 | ↑ higher | Source ↗official | |
| SALAD-Benchattack_enhanced_information_safety_harms | #8 / 33 | 27.04 | ↑ higher | Source ↗official | |
| SALAD-Benchattack_enhanced_malicious_use | #9 / 33 | 23.82 | ↑ higher | Source ↗official | |
| SALAD-Benchattack_enhanced_misinformation_harms | #10 / 33 | 24.84 | ↑ higher | Source ↗official | |
| SALAD-Benchattack_enhanced_representation_toxicity | #12 / 33 | 20.24 | ↑ higher | Source ↗official | |
| SALAD-Benchattack_enhanced_socioeconomic_harms | #10 / 33 | 23.38 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_human_autonomy_integrity | #25 / 33 | 86.78 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_information_safety_harms | #20 / 33 | 93.77 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_malicious_use | #26 / 33 | 82.67 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_misinformation_harms | #27 / 33 | 86.12 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_representation_toxicity | #27 / 33 | 78.37 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_socioeconomic_harms | #20 / 33 | 86.72 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_human_autonomy_integrity | #28 / 33 | 5.833 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_information_safety_harms | #28 / 33 | 3.889 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_malicious_use | #27 / 33 | 5.962 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_misinformation_harms | #28 / 33 | 5.238 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_representation_toxicity | #27 / 33 | 4.792 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_socioeconomic_harms | #28 / 33 | 6.667 | ↑ higher | Source ↗official |