Model profile
Baichuan 2 13B Chat
Baichuandeveloper
2023-09-06release date
#148 / 309overall rank
13eval lineages
Evidence summary
Baichuan 2 13B Chat has an estimated overall rank of #148; its 90% source-sensitivity interval is #71–#210. Its behavior-only rank is #154; company governance moves the combined estimate to #148. Published evidence spans 13 evals and 5 of 7 behavior components. Its strongest relative result is SuperCLUE Safety (traditional_safety, #1 of 31); its weakest is SafetyBench (UB, #20 of 21).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Hugging Face ↗baichuan-inc/Baichuan2-13B-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 | Source |
|---|---|---|---|---|
| ChineseSafescore | #1 / 22 | ↑70.43 | Source ↗official | |
| ChiSafetyBenchharmful_response_rate | #5 / 14 | ↓0.43 | Source ↗official | |
| ChiSafetyBenchmcq_score | #7 / 12 | ↑79.83 | Source ↗official | |
| CMoralEvalfamilial_morality | #13 / 26 | ↑0.39 | Source ↗official | |
| CMoralEvalinternet_ethics | #19 / 26 | ↑0.34 | Source ↗official | |
| CMoralEvalpersonal_morality | #17 / 26 | ↑0.36 | Source ↗official | |
| CMoralEvalprofessional_ethics | #19 / 26 | ↑0.36 | Source ↗official | |
| CMoralEvalsocial_morality | #16 / 26 | ↑0.37 | Source ↗official | |
| CRiskEvaldeception_willingness | #15 / 17 | ↓37.75 | Source ↗official | |
| CRiskEvaldesire_for_resource | #11 / 17 | ↓35.35 | Source ↗official | |
| CRiskEvalharmful_goal | #15 / 17 | ↓54.94 | Source ↗official | |
| CRiskEvalimprovement_intent | #10 / 17 | ↓50.82 | Source ↗official | |
| CRiskEvalmalicious_coordination | #12 / 17 | ↓18.46 | Source ↗official | |
| CRiskEvalself_preservation | #15 / 17 | ↓43.45 | Source ↗official | |
| CRiskEvalsituational_awareness | #7 / 17 | ↓63.34 | Source ↗official | |
| Fake Alignment (FINE)multiple_choice_safe_decision_rate | #9 / 14 | ↑45.33 | Source ↗official | |
| Fake Alignment (FINE)open_ended_safe_response_rate | #1 / 14 | ↑100 | Source ↗official | |
| FinEval Financial Security Knowledgefinancial_security_accuracy_pct | #16 / 19 | ↑61.6 | Source ↗official | |
| FLAMESdata_protection | #9 / 13 | ↑39.47 | Source ↗official | |
| FLAMESfairness | #6 / 13 | ↑38.55 | Source ↗official | |
| FLAMESlegality | #7 / 13 | ↑39.13 | Source ↗official | |
| FLAMESmorality | #6 / 13 | ↑44.78 | Source ↗official | |
| FLAMESsafety | #2 / 13 | ↑53.85 | Source ↗official | |
| HarmBenchdr | #14 / 28 | ↓19.3 | Source ↗official | |
| M3-SafetyBenchoverall_score | #13 / 19 | ↑86.39 | Source ↗official | |
| S-Evalbase_en_overall | #4 / 22 | ↑77.4 | Source ↗official | |
| SafeDialBenchaggression | #12 / 18 | ↑7.03 | Source ↗official | |
| SafeDialBenchethics | #7 / 18 | ↑7.6 | Source ↗official | |
| SafeDialBenchfairness | #15 / 18 | ↑7.17 | Source ↗official | |
| SafeDialBenchlegality | #4 / 18 | ↑7.987 | Source ↗official | |
| SafeDialBenchmorality | #6 / 18 | ↑7.303 | Source ↗official | |
| SafeDialBenchprivacy | #5 / 18 | ↑7.617 | Source ↗official | |
| SafetyBenchEM | #5 / 21 | ↑75.75 | Source ↗official | |
| SafetyBenchIA | #4 / 21 | ↑82.65 | Source ↗official | |
| SafetyBenchMH | #7 / 21 | ↑83.65 | Source ↗official | |
| SafetyBenchOFF | #6 / 21 | ↑69.25 | Source ↗official | |
| SafetyBenchPH | #5 / 21 | ↑76.35 | Source ↗official | |
| SafetyBenchPP | #4 / 21 | ↑82.05 | Source ↗official | |
| SafetyBenchUB | #20 / 21 | ↑49.2 | Source ↗official | |
| SuperCLUE Safetyinstruction_attack | #3 / 31 | ↑72.41 | Source ↗official | |
| SuperCLUE Safetyresponsible_ai | #4 / 31 | ↑72.73 | Source ↗official | |
| SuperCLUE Safetytraditional_safety | #1 / 31 | ↑87.23 | Source ↗official |