Model profile
Vicuna 7B
LMSYSdeveloper
2023-06-22release date
#274 / 312overall rank
4eval lineages
Evidence summary
Vicuna 7B has an estimated overall rank of #274; its 90% source-sensitivity interval is #198–#285. Its behavior-only rank is #280; company governance moves the combined estimate to #274. Published evidence spans 4 evals and 3 of 7 behavior components. Its strongest relative result is SafetyBench (OFF, #14 of 21); its weakest is SafetyBench (UB, #19 of 21).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Hugging Face ↗lmsys/vicuna-7b-v1.3
- 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 |
|---|---|---|---|---|
| Do-Not-Answerhuman_harmlessness_rate | #5 / 6 | ↑94.46 | Source ↗official | |
| Fake Alignment (FINE)multiple_choice_safe_decision_rate | #11 / 14 | ↑25.33 | Source ↗official | |
| Fake Alignment (FINE)open_ended_safe_response_rate | #12 / 14 | ↑89.33 | Source ↗official | |
| S-Evalbase_en_overall | #17 / 22 | ↑50.2 | Source ↗official | |
| SafetyBenchEM | #15 / 21 | ↑59.8 | Source ↗official | |
| SafetyBenchIA | #18 / 21 | ↑65.1 | Source ↗official | |
| SafetyBenchMH | #16 / 21 | ↑73.1 | Source ↗official | |
| SafetyBenchOFF | #14 / 21 | ↑65.1 | Source ↗official | |
| SafetyBenchPH | #15 / 21 | ↑60.9 | Source ↗official | |
| SafetyBenchPP | #17 / 21 | ↑68.4 | Source ↗official | |
| SafetyBenchUB | #19 / 21 | ↑52.7 | Source ↗official |