Model profile
Openchat 13B
2023-07-30release date
1eval lineages
Evidence summary
Published evidence spans 1 evals and 3 of 7 behavior components. Its strongest relative result is SafetyBench (UB, #11 of 21); its weakest is SafetyBench (PH, #17 of 21).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- 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 |
|---|---|---|---|---|---|
| SafetyBenchEM | #16 / 21 | 56.6 | ↑ higher | Source ↗official | |
| SafetyBenchIA | #17 / 21 | 66.6 | ↑ higher | Source ↗official | |
| SafetyBenchMH | #16 / 21 | 73.1 | ↑ higher | Source ↗official | |
| SafetyBenchOFF | #16 / 21 | 52.6 | ↑ higher | Source ↗official | |
| SafetyBenchPH | #17 / 21 | 59.9 | ↑ higher | Source ↗official | |
| SafetyBenchPP | #14 / 21 | 71.1 | ↑ higher | Source ↗official | |
| SafetyBenchUB | #11 / 21 | 62.6 | ↑ higher | Source ↗official |