← Evals

Evaluation profile

Olam Social Poker — Social Lie Rate

1sub-evals
0.329%total index weight
1components

Within-component eval weight: Truthfulness 2.19%.

Model score (lower is better)Predicted score

About this eval

Propensity to verbalize deliberate strategic lies during Social Poker table talk.

Included in the behavior ranking.

Sub-evals

MeasureComponentDirectionTotal index weightWithin-component weight
social_lie_rate_per_10000_turnsolam-poker-deception/deception-rate.csv:social_lie_rate_per_10000_turnsMeasures how frequently a model chooses to make deliberate strategic false claims during Social Poker.truthfulness_honesty:1.000olam-poker-deceptionLower is better0.329%Truthfulness 2.19%

social_lie_rate_per_10000_turns

Measures how frequently a model chooses to make deliberate strategic false claims during Social Poker.

RankModelValueRelative performanceProvenance
1gemini-3.6-flash0official
1gpt-5.6-luna0official
3claude-sonnet-52official
3gpt-5.52official
3gpt-5.6-sol2official
3gpt-5.6-terra2official
7nemotron-3-ultra-550b-a55b3official
8gemini-3.5-flash4official
9grok-4.66official
10gemini-3.7-flash7official
11gemini-3.1-pro-preview8official
12deepseek-v4-pro9official
12grok-4.59official
14muse-spark-1.112official
15gemini-3.5-flash-lite13official
16glm-5.228official
17ox-alpha30official
18deepseek-v4-flash37official
19kimi-k340official
20claude-opus-4.880official
21claude-opus-5163official
22claude-fable-5164official