← Models

Model profile

Qwen2.5 7B

Alibabadeveloper
2024-09-19release date
#300 / 331Safety rank
#444 / 631Freedom rank

Evidence summary

Safety. Qwen2.5 7B has an estimated Safety rank of #300; its 90% source-sensitivity interval is #241–#303. Its behavior-only rank is #298; company governance moves the combined estimate to #300. Published Safety evidence spans 3 eval lineages and 3 of 7 components. Its strongest relative result is IndoBias-Pairs — parity-aware culturally grounded bias (parity_score, #15 of 26); its weakest is HalluVerse-M3 Hallucination Recognition (hallucination_recognition_accuracy, #13 of 14).

Freedom. Qwen2.5 7B has an estimated Freedom rank of #444; its 90% source-sensitivity interval is #260–#575. Published Freedom evidence spans 1 eval lineages and 1 of 1 components. Its strongest relative result is VETO Misfired Alignment (misfired_alignment_rate_pct, #16 of 25); its weakest is VETO Misfired Alignment (misfired_alignment_rate_pct, #16 of 25).

Compare this model

Only models sharing at least one published sub-eval are listed.

Official and reference links

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-evalRankValueDistributionSource
HalluVerse-M3 Hallucination Recognitionhallucination_recognition_accuracy#13 / 1459.4Source ↗official
IndoBias-Pairs — parity-aware culturally grounded biasparity_score#15 / 2685.28Source ↗official
VETO Misfired Alignmentmisfired_alignment_rate_pct#16 / 258.9Source ↗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-evalRankValueDistributionSource
VETO Misfired Alignmentmisfired_alignment_rate_pct#16 / 258.9Source ↗official