Model profile
Qwen1.5 7B Chat
Alibabadeveloper
2024-02-04release date
#180 / 305overall rank
9eval lineages
Evidence summary
Qwen1.5 7B Chat has an estimated overall rank of #180; its 90% source-sensitivity interval is #110–#224. Its behavior-only rank is #168; company governance moves the combined estimate to #180. Published evidence spans 9 evals and 6 of 7 behavior components. Its strongest relative result is CRiskEval (malicious_coordination, #1 of 17); its weakest is SALAD-Bench (attack_enhanced_socioeconomic_harms, #27 of 33).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Hugging Face ↗Qwen/Qwen1.5-7B-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 | #6 / 22 | ↑62.48 | Source ↗official | |
| ChiSafetyBenchharmful_response_rate | #5 / 14 | ↓0.43 | Source ↗official | |
| ChiSafetyBenchmcq_score | #8 / 12 | ↑79.39 | Source ↗official | |
| Contextual MoralChoicehuman_agreement | #12 / 22 | ↑0.4 | Source ↗official | |
| CRiskEvaldeception_willingness | #9 / 17 | ↓22.6 | Source ↗official | |
| CRiskEvaldesire_for_resource | #8 / 17 | ↓32.14 | Source ↗official | |
| CRiskEvalharmful_goal | #9 / 17 | ↓37.92 | Source ↗official | |
| CRiskEvalimprovement_intent | #6 / 17 | ↓47.47 | Source ↗official | |
| CRiskEvalmalicious_coordination | #1 / 17 | ↓5.68 | Source ↗official | |
| CRiskEvalself_preservation | #7 / 17 | ↓36.01 | Source ↗official | |
| CRiskEvalsituational_awareness | #13 / 17 | ↓67.2 | Source ↗official | |
| HyperCLOVA X Toxic Continuation Panelskold_toxic_count | #2 / 7 | ↓0.0036 | Source ↗official | |
| HyperCLOVA X Toxic Continuation Panelskold_toxicity | #2 / 7 | ↓0.1061 | Source ↗official | |
| HyperCLOVA X Toxic Continuation Panelsrealt-toxicprompts_toxic_count | #3 / 7 | ↓0.006 | Source ↗official | |
| HyperCLOVA X Toxic Continuation Panelsrealt-toxicprompts_toxicity | #3 / 7 | ↓0.055 | Source ↗official | |
| JailBenchjailbreak_success_rate | #11 / 14 | ↓71.6 | Source ↗official | |
| OR-Benchover_refusal_rate | #10 / 25 | ↓39.2 | Source ↗official | |
| OR-Benchtoxic_acceptance_rate | #18 / 25 | ↓15 | Source ↗official | |
| SALAD-Benchattack_enhanced_human_autonomy_integrity | #22 / 33 | ↑9.91 | Source ↗official | |
| SALAD-Benchattack_enhanced_information_safety_harms | #21 / 33 | ↑7.17 | Source ↗official | |
| SALAD-Benchattack_enhanced_malicious_use | #23 / 33 | ↑8.32 | Source ↗official | |
| SALAD-Benchattack_enhanced_misinformation_harms | #19 / 33 | ↑9.38 | Source ↗official | |
| SALAD-Benchattack_enhanced_representation_toxicity | #23 / 33 | ↑7.39 | Source ↗official | |
| SALAD-Benchattack_enhanced_socioeconomic_harms | #27 / 33 | ↑7.36 | Source ↗official | |
| SALAD-Benchbase_human_autonomy_integrity | #16 / 33 | ↑95.17 | Source ↗official | |
| SALAD-Benchbase_information_safety_harms | #14 / 33 | ↑95.67 | Source ↗official | |
| SALAD-Benchbase_malicious_use | #16 / 33 | ↑94.12 | Source ↗official | |
| SALAD-Benchbase_misinformation_harms | #14 / 33 | ↑94.04 | Source ↗official | |
| SALAD-Benchbase_representation_toxicity | #15 / 33 | ↑90.23 | Source ↗official | |
| SALAD-Benchbase_socioeconomic_harms | #13 / 33 | ↑91.19 | Source ↗official | |
| SALAD-Benchmcq_human_autonomy_integrity | #24 / 33 | ↑19.17 | Source ↗official | |
| SALAD-Benchmcq_information_safety_harms | #26 / 33 | ↑11.94 | Source ↗official | |
| SALAD-Benchmcq_malicious_use | #24 / 33 | ↑16.35 | Source ↗official | |
| SALAD-Benchmcq_misinformation_harms | #24 / 33 | ↑18.81 | Source ↗official | |
| SALAD-Benchmcq_representation_toxicity | #25 / 33 | ↑15.62 | Source ↗official | |
| SALAD-Benchmcq_socioeconomic_harms | #24 / 33 | ↑17.78 | Source ↗official | |
| SORRY-Benchavg | #39 / 51 | ↓0.4 | Source ↗official |