Model profile
Vicuna 13B V1.5
LMSYSdeveloper
2023-03-30release date
#222 / 267overall rank
4eval lineages
Evidence summary
Vicuna 13B V1.5 has an estimated overall rank of #222; its 90% source-sensitivity interval is #146–#237. Its behavior-only rank is #226; company governance moves the combined estimate to #222. Published evidence spans 4 evals and 4 of 7 behavior components. Its strongest relative result is HarmBench (dr, #15 of 28); its weakest is Large-scale Moral Machine experiment on LLMs (human_choice_distance, #31 of 39).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Hugging Face ↗lmsys/vicuna-13b-v1.5
- Official model page ↗Exact model document · Reviewed official Hugging Face owner · official repository
- Release source ↗global scoring family inherited
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 |
|---|---|---|---|---|---|
| HarmBenchdr | #15 / 28 | 19.8 | ↓ lower | Source ↗official | |
| JailBenchjailbreak_success_rate | #9 / 14 | 66.32 | ↓ lower | Source ↗official | |
| Large-scale Moral Machine experiment on LLMshuman_choice_distance | #31 / 39 | 1.256 | ↓ lower | Source ↗official | |
| SORRY-Benchavg | #28 / 51 | 0.32 | ↓ lower | Source ↗official |