Model profile
Qwen1.5 4B Chat
2024-02-04release date
2eval lineages
Evidence summary
Published evidence spans 2 evals and 4 of 7 behavior components. Its strongest relative result is SALAD-Bench (base_misinformation_harms, #7 of 33); its weakest is SALAD-Bench (mcq_representation_toxicity, #31 of 33).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Hugging Face ↗Qwen/Qwen1.5-4B-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 |
|---|---|---|---|---|---|
| SALAD-Benchattack_enhanced_human_autonomy_integrity | #8 / 33 | 26.29 | ↑ higher | Source ↗official | |
| SALAD-Benchattack_enhanced_information_safety_harms | #10 / 33 | 23.13 | ↑ higher | Source ↗official | |
| SALAD-Benchattack_enhanced_malicious_use | #10 / 33 | 22.84 | ↑ higher | Source ↗official | |
| SALAD-Benchattack_enhanced_misinformation_harms | #7 / 33 | 28.29 | ↑ higher | Source ↗official | |
| SALAD-Benchattack_enhanced_representation_toxicity | #11 / 33 | 22.09 | ↑ higher | Source ↗official | |
| SALAD-Benchattack_enhanced_socioeconomic_harms | #13 / 33 | 19.05 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_human_autonomy_integrity | #11 / 33 | 96.62 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_information_safety_harms | #13 / 33 | 96.14 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_malicious_use | #9 / 33 | 96.93 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_misinformation_harms | #7 / 33 | 96.45 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_representation_toxicity | #10 / 33 | 93.05 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_socioeconomic_harms | #7 / 33 | 94.01 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_human_autonomy_integrity | #30 / 33 | 0 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_information_safety_harms | #30 / 33 | 0 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_malicious_use | #31 / 33 | 0 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_misinformation_harms | #30 / 33 | 0 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_representation_toxicity | #31 / 33 | 0 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_socioeconomic_harms | #30 / 33 | 0 | ↑ higher | Source ↗official | |
| SORRY-Benchavg | #24 / 51 | 0.27 | ↓ lower | Source ↗official |