Model profile
ChatGLM 6B
Evidence summary
Safety. ChatGLM 6B has an estimated Safety rank of #260; its 90% source-sensitivity interval is #128–#291. Its behavior-only rank is #261; company governance moves the combined estimate to #260. Published Safety evidence spans 5 eval lineages and 4 of 7 components. Its strongest relative result is SafetyPrompts (score, #3 of 15); its weakest is ChiSafetyBench (harmful_response_rate, #14 of 14).
Freedom. ChatGLM 6B has an estimated Freedom rank of #141; its 90% source-sensitivity interval is #89–#340. Published Freedom evidence spans 3 eval lineages and 1 of 1 components. Its strongest relative result is ChiSafetyBench (harmful_response_rate, #1 of 14); its weakest is SafetyPrompts (score, #13 of 15).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Hugging Face ↗zai-org/chatglm-6b
- Official model page ↗Exact model document · Reviewed official Hugging Face owner · official repository
- Release source ↗direct research preview identity date
Safety evals
Rank and direction are specific to the Safety portfolio. Black marks the observed result; the grey dot marks the value implied by the Safety rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| Chinese Bias Benchmark for Question Answeringbias_score | #7 / 10 | ↓0.4437 | Source ↗official | |
| ChiSafetyBenchharmful_response_rate | #14 / 14 | ↓1.95 | Source ↗official | |
| CValuesscore | #3 / 10 | ↑95 | Source ↗official | |
| FLAMESdata_protection | #11 / 13 | ↑32.89 | Source ↗official | |
| FLAMESfairness | #12 / 13 | ↑26.91 | Source ↗official | |
| FLAMESlegality | #5 / 13 | ↑50 | Source ↗official | |
| FLAMESmorality | #9 / 13 | ↑40.3 | Source ↗official | |
| FLAMESsafety | #12 / 13 | ↑15.38 | Source ↗official | |
| SafetyPromptsscore | #3 / 15 | ↑96.81 | Source ↗official |
Freedom evals
Rank and direction are specific to the Freedom portfolio. Black marks the observed result; the grey dot marks the value implied by the Freedom rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| ChiSafetyBenchharmful_response_rate | #1 / 14 | ↑1.95 | Source ↗official | |
| ChiSafetyBenchrefusal_rr1 | #6 / 14 | ↓72.08 | Source ↗official | |
| ChiSafetyBenchrefusal_rr2 | #6 / 14 | ↓72.08 | Source ↗official | |
| FLAMESdata_protection | #2 / 13 | ↓32.89 | Source ↗official | |
| FLAMESfairness | #2 / 13 | ↓26.91 | Source ↗official | |
| FLAMESlegality | #8 / 13 | ↓50 | Source ↗official | |
| FLAMESmorality | #4 / 13 | ↓40.3 | Source ↗official | |
| FLAMESsafety | #1 / 13 | ↓15.38 | Source ↗official | |
| SafetyPromptsscore | #13 / 15 | ↓96.81 | Source ↗official |
Values evaluations
Descriptive values and political-framing results are separate from safety/ethics ranks. Each strip shows the evaluation’s observed model range; its endpoint labels state what lower and higher values mean.