qev-450m-mlx / reports /paired_tests.json
twainsk's picture
Publish Qev v0.4.0 qev-450m with unchanged LFM2.5 foundation
2aa0473 verified
Raw History Blame Contribute Delete
1.9 kB
[
{
"a": "vl-a",
"b": "qwen35-causal",
"subset": "general",
"n": 880,
"accuracy_a": 0.7886363636363637,
"accuracy_b": 0.8170454545454545,
"only_a_correct": 44,
"only_b_correct": 69,
"exact_mcnemar_p": 0.023534002126010038
},
{
"a": "vl-a",
"b": "qwen35-causal",
"subset": "snake",
"n": 500,
"accuracy_a": 0.988,
"accuracy_b": 0.992,
"only_a_correct": 1,
"only_b_correct": 3,
"exact_mcnemar_p": 0.625
},
{
"a": "t230-a",
"b": "qwen35-causal",
"subset": "general",
"n": 880,
"accuracy_a": 0.7738636363636363,
"accuracy_b": 0.8170454545454545,
"only_a_correct": 52,
"only_b_correct": 90,
"exact_mcnemar_p": 0.001801710219750172
},
{
"a": "t230-a",
"b": "qwen35-causal",
"subset": "snake",
"n": 500,
"accuracy_a": 0.99,
"accuracy_b": 0.992,
"only_a_correct": 2,
"only_b_correct": 3,
"exact_mcnemar_p": 1.0
},
{
"a": "vl-a",
"b": "t230-a",
"subset": "general",
"n": 880,
"accuracy_a": 0.7886363636363637,
"accuracy_b": 0.7738636363636363,
"only_a_correct": 75,
"only_b_correct": 62,
"exact_mcnemar_p": 0.3052427781333533
},
{
"a": "vl-a",
"b": "t230-a",
"subset": "snake",
"n": 500,
"accuracy_a": 0.988,
"accuracy_b": 0.99,
"only_a_correct": 2,
"only_b_correct": 3,
"exact_mcnemar_p": 1.0
},
{
"a": "vl-a",
"b": "lfm2-causal",
"subset": "general",
"n": 880,
"accuracy_a": 0.7886363636363637,
"accuracy_b": 0.7568181818181818,
"only_a_correct": 73,
"only_b_correct": 45,
"exact_mcnemar_p": 0.012602628706921444
},
{
"a": "vl-a",
"b": "lfm2-causal",
"subset": "snake",
"n": 500,
"accuracy_a": 0.988,
"accuracy_b": 0.986,
"only_a_correct": 3,
"only_b_correct": 2,
"exact_mcnemar_p": 1.0
}
]