`torch_dtype` is deprecated! Use `dtype` instead!
easy: 25/48

== easy: n=48 acc=0.667 cci=53.4 brier=0.5289 ece=0.1662 p50=0.033s p95=0.039s
   extraction             9/12 = 0.750
   fact                   10/12 = 0.833
   intent                 5/12 = 0.417
   tool_selection         8/12 = 0.667
hard: 25/111
hard: 50/111
hard: 75/111
hard: 100/111

== hard: n=111 acc=0.342 cci=0.9 brier=0.8243 ece=0.2794 p50=0.036s p95=0.099s
   adversarial            3/6 = 0.500
   ambiguous              2/7 = 0.286
   judge_hard             8/17 = 0.471
   long_policy            6/19 = 0.316
   multi_hop              4/18 = 0.222
   probability            5/10 = 0.500
   routing_hard           3/5 = 0.600
   temporal_numeric       3/15 = 0.200
   tradeoff               2/6 = 0.333
   trap                   2/8 = 0.250
original: 25/72
original: 50/72

== original: n=72 acc=0.403 cci=13.3 brier=0.7913 ece=0.2987 p50=0.033s p95=0.034s
   adequacy               4/12 = 0.333
   extraction             5/12 = 0.417
   intent                 5/12 = 0.417
   ordinal                4/12 = 0.333
   policy                 6/12 = 0.500
   routing                5/12 = 0.417
wrote /root/alea/jevbench_results.json