Model profile
Llama 3.1 Tulu 3 70B Sft
2024-11-18release date
1eval lineages
Evidence summary
Published evidence spans 1 evals and 2 of 7 behavior components. Its strongest relative result is AbstentionBench (subjective_f1, #3 of 20); its weakest is AbstentionBench (underspecified_context_f1, #12 of 20).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Hugging Face ↗allenai/Llama-3.1-Tulu-3-70B-SFT
- Official model page ↗Exact model document · Reviewed official Hugging Face owner · official repository
- Release source ↗repository created at proxy
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 |
|---|---|---|---|---|---|
| AbstentionBenchanswer_unknown_f1 | #6 / 20 | 0.8931 | ↑ higher | Source ↗official | |
| AbstentionBenchfalse_premise_f1 | #8 / 20 | 0.7105 | ↑ higher | Source ↗official | |
| AbstentionBenchstale_f1 | #6 / 20 | 0.6742 | ↑ higher | Source ↗official | |
| AbstentionBenchsubjective_f1 | #3 / 20 | 0.7804 | ↑ higher | Source ↗official | |
| AbstentionBenchunderspecified_context_f1 | #12 / 20 | 0.5215 | ↑ higher | Source ↗official | |
| AbstentionBenchunderspecified_intent_f1 | #7 / 20 | 0.7492 | ↑ higher | Source ↗official |