Model profile
Flan T5 Xxl
2022-10-21release date
1eval lineages
Evidence summary
Published evidence spans 1 evals and 3 of 7 behavior components. Its strongest relative result is SafetyBench (OFF, #2 of 21); its weakest is SafetyBench (PH, #13 of 21).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Hugging Face ↗google/flan-t5-xxl
- 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 |
|---|---|---|---|---|---|
| SafetyBenchEM | #9 / 21 | 69.5 | ↑ higher | Source ↗official | |
| SafetyBenchIA | #10 / 21 | 78.2 | ↑ higher | Source ↗official | |
| SafetyBenchMH | #13 / 21 | 77.9 | ↑ higher | Source ↗official | |
| SafetyBenchOFF | #2 / 21 | 79.2 | ↑ higher | Source ↗official | |
| SafetyBenchPH | #13 / 21 | 67 | ↑ higher | Source ↗official | |
| SafetyBenchPP | #10 / 21 | 76.4 | ↑ higher | Source ↗official | |
| SafetyBenchUB | #2 / 21 | 70.2 | ↑ higher | Source ↗official |