Model profile
Evidence summary
Safety. Tulu 2 DPO 13B does not meet the evidence gate for a Safety rank. Published Safety evidence spans 1 eval lineages and 4 of 7 components. Its strongest relative result is SALAD-Bench (mcq_representation_toxicity, #12 of 33); its weakest is SALAD-Bench (attack_enhanced_human_autonomy_integrity, #31 of 33).
Freedom. Tulu 2 DPO 13B has an estimated Freedom rank of #310; its 90% source-sensitivity interval is #83–#627. Published Freedom evidence spans 1 eval lineages and 1 of 1 components. Its strongest relative result is SALAD-Bench (attack_enhanced_human_autonomy_integrity, #2 of 33); its weakest is SALAD-Bench (base_information_safety_harms, #12 of 33).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Hugging Face ↗allenai/tulu-2-dpo-13b
- Official model page ↗Exact model document · Reviewed official Hugging Face owner · official repository
- Release source ↗direct research preview identity date
Safety evals
Rank and direction are specific to the Safety portfolio. Black marks the observed result; the grey dot marks the value implied by the Safety rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| SALAD-Benchattack_enhanced_human_autonomy_integrity | #31 / 33 | ↑4.53 | Source ↗official | |
| SALAD-Benchattack_enhanced_information_safety_harms | #28 / 33 | ↑4.56 | Source ↗official | |
| SALAD-Benchattack_enhanced_malicious_use | #27 / 33 | ↑6.12 | Source ↗official | |
| SALAD-Benchattack_enhanced_misinformation_harms | #28 / 33 | ↑5.76 | Source ↗official | |
| SALAD-Benchattack_enhanced_representation_toxicity | #28 / 33 | ↑5.59 | Source ↗official | |
| SALAD-Benchattack_enhanced_socioeconomic_harms | #29 / 33 | ↑6.93 | Source ↗official | |
| SALAD-Benchbase_human_autonomy_integrity | #23 / 33 | ↑89.52 | Source ↗official | |
| SALAD-Benchbase_information_safety_harms | #22 / 33 | ↑92.82 | Source ↗official | |
| SALAD-Benchbase_malicious_use | #24 / 33 | ↑85.52 | Source ↗official | |
| SALAD-Benchbase_misinformation_harms | #24 / 33 | ↑88.48 | Source ↗official | |
| SALAD-Benchbase_representation_toxicity | #23 / 33 | ↑85.28 | Source ↗official | |
| SALAD-Benchbase_socioeconomic_harms | #23 / 33 | ↑84.49 | Source ↗official | |
| SALAD-Benchmcq_human_autonomy_integrity | #14 / 33 | ↑41.67 | Source ↗official | |
| SALAD-Benchmcq_information_safety_harms | #16 / 33 | ↑34.44 | Source ↗official | |
| SALAD-Benchmcq_malicious_use | #15 / 33 | ↑40.13 | Source ↗official | |
| SALAD-Benchmcq_misinformation_harms | #12 / 33 | ↑40.71 | Source ↗official | |
| SALAD-Benchmcq_representation_toxicity | #12 / 33 | ↑40.31 | Source ↗official | |
| SALAD-Benchmcq_socioeconomic_harms | #14 / 33 | ↑38.89 | Source ↗official |
Freedom evals
Rank and direction are specific to the Freedom portfolio. Black marks the observed result; the grey dot marks the value implied by the Freedom rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| SALAD-Benchattack_enhanced_human_autonomy_integrity | #2 / 33 | ↓4.53 | Source ↗official | |
| SALAD-Benchattack_enhanced_information_safety_harms | #6 / 33 | ↓4.56 | Source ↗official | |
| SALAD-Benchattack_enhanced_malicious_use | #7 / 33 | ↓6.12 | Source ↗official | |
| SALAD-Benchattack_enhanced_misinformation_harms | #6 / 33 | ↓5.76 | Source ↗official | |
| SALAD-Benchattack_enhanced_representation_toxicity | #6 / 33 | ↓5.59 | Source ↗official | |
| SALAD-Benchattack_enhanced_socioeconomic_harms | #4 / 33 | ↓6.93 | Source ↗official | |
| SALAD-Benchbase_human_autonomy_integrity | #11 / 33 | ↓89.52 | Source ↗official | |
| SALAD-Benchbase_information_safety_harms | #12 / 33 | ↓92.82 | Source ↗official | |
| SALAD-Benchbase_malicious_use | #10 / 33 | ↓85.52 | Source ↗official | |
| SALAD-Benchbase_misinformation_harms | #10 / 33 | ↓88.48 | Source ↗official | |
| SALAD-Benchbase_representation_toxicity | #11 / 33 | ↓85.28 | Source ↗official | |
| SALAD-Benchbase_socioeconomic_harms | #11 / 33 | ↓84.49 | Source ↗official |