Model profile
Yi 1.5 9B Chat
Evidence summary
Safety. Yi 1.5 9B Chat has an estimated Safety rank of #255; its 90% source-sensitivity interval is #108–#305. Its behavior-only rank is #256; company governance moves the combined estimate to #255. Published Safety evidence spans 3 eval lineages and 3 of 7 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (toxicity_attack_non_success_rate, #40 of 270); its weakest is Enkrypt AI Safety Leaderboard (harmful_attack_non_success_rate, #235 of 272).
Freedom. Yi 1.5 9B Chat has an estimated Freedom rank of #533; its 90% source-sensitivity interval is #342–#662. Published Freedom evidence spans 1 eval lineages and 1 of 1 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (harmful_attack_non_success_rate, #37 of 272); its weakest is Enkrypt AI Safety Leaderboard (toxicity_attack_non_success_rate, #230 of 270).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- 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 |
|---|---|---|---|---|
| ChineseSafescore | #8 / 22 | ↑60.35 | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #160 / 272 | ↑14.73 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #108 / 272 | ↑88.83 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #235 / 272 | ↑40.56 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #40 / 270 | ↑99 | Source ↗official | |
| FinEval Financial Security Knowledgefinancial_security_accuracy_pct | #13 / 19 | ↑71.4 | 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 |
|---|---|---|---|---|
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #164 / 272 | ↓88.83 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #37 / 272 | ↓40.56 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #230 / 270 | ↓99 | Source ↗official |