Model profile
Moss Moon 003 Sft
Fudan NLPdeveloper
2023-04-19release date
#242 / 267overall rank
3eval lineages
Evidence summary
Moss Moon 003 Sft has an estimated overall rank of #242; its 90% source-sensitivity interval is #183–#255. Its behavior-only rank is #248; company governance moves the combined estimate to #242. Published evidence spans 3 evals and 3 of 7 behavior components. Its strongest relative result is CMoralEval (social_morality, #14 of 26); its weakest is SuperCLUE Safety (traditional_safety, #31 of 31).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Hugging Face ↗OpenMOSS-Team/moss-moon-003-sft
- 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 |
|---|---|---|---|---|---|
| CMoralEvalfamilial_morality | #17 / 26 | 0.38 | ↑ higher | Source ↗official | |
| CMoralEvalinternet_ethics | #14 / 26 | 0.39 | ↑ higher | Source ↗official | |
| CMoralEvalpersonal_morality | #14 / 26 | 0.39 | ↑ higher | Source ↗official | |
| CMoralEvalprofessional_ethics | #15 / 26 | 0.38 | ↑ higher | Source ↗official | |
| CMoralEvalsocial_morality | #14 / 26 | 0.39 | ↑ higher | Source ↗official | |
| Fake Alignment (FINE)multiple_choice_safe_decision_rate | #14 / 14 | 10.67 | ↑ higher | Source ↗official | |
| Fake Alignment (FINE)open_ended_safe_response_rate | #9 / 14 | 94.67 | ↑ higher | Source ↗official | |
| SuperCLUE Safetyinstruction_attack | #31 / 31 | 44.83 | ↑ higher | Source ↗official | |
| SuperCLUE Safetyresponsible_ai | #29 / 31 | 43.64 | ↑ higher | Source ↗official | |
| SuperCLUE Safetytraditional_safety | #31 / 31 | 63.83 | ↑ higher | Source ↗official |