← Models

Model profile

Palmyra X 004

Writerdeveloper
2024-10-09release date
#79 / 346Safety rank
#169 / 662Freedom rank

Evidence summary

Safety. Palmyra X 004 has an estimated Safety rank of #79; its 90% source-sensitivity interval is #41–#334. Its behavior-only rank is #76; company governance moves the combined estimate to #79. Published Safety evidence spans 6 eval lineages and 3 of 7 components. Its strongest relative result is HELM Safety (simple_safety_tests, #1 of 80); its weakest is AIRBench 2024 Safety Scenarios (safety_scenarios, #70 of 80).

Freedom. Palmyra X 004 has an estimated Freedom rank of #169; its 90% source-sensitivity interval is #32–#566. Published Freedom evidence spans 5 eval lineages and 1 of 1 components. Its strongest relative result is HELM Safety (xstest, #6 of 80); its weakest is HELM Safety (simple_safety_tests, #58 of 80).

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
AIRBench 2024 Safety Scenariossafety_scenarios#70 / 80↑0.442Source ↗official
HELM Safetyanthropic_red_team#47 / 80↑0.986Source ↗official
HELM Safetybbq#22 / 80↑0.955Source ↗official
HELM Safetyharmbench#40 / 80↑0.738Source ↗official
HELM Safetysimple_safety_tests#1 / 80↑1Source ↗official
HELM Safetyxstest#6 / 80↑0.984Source ↗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
AIRBench 2024 Safety Scenariossafety_scenarios#11 / 80↓0.442Source ↗official
HELM Safetyanthropic_red_team#33 / 80↓0.986Source ↗official
HELM Safetyharmbench#40 / 80↓0.738Source ↗official
HELM Safetysimple_safety_tests#58 / 80↓1Source ↗official
HELM Safetyxstest#6 / 80↑0.984Source ↗official