MMMMoK9922's picture
Upload candidate_value_qslow_vlm model and summary
f7031b4 verified
Raw History Blame Contribute Delete
1.42 kB
{
"heads": 4,
"test": [
{
"groups": 36,
"selected_successes": 30,
"slow_successes": 35,
"oracle_successes": 36,
"preferred_accuracy": 0.8333333333333334,
"success_accuracy": 0.9074074074074074,
"selected_speed_counts": {
"fast": 36,
"normal": 0,
"slow": 0
}
},
{
"groups": 36,
"selected_successes": 30,
"slow_successes": 35,
"oracle_successes": 36,
"preferred_accuracy": 0.8333333333333334,
"success_accuracy": 0.9074074074074074,
"selected_speed_counts": {
"fast": 36,
"normal": 0,
"slow": 0
}
},
{
"groups": 36,
"selected_successes": 30,
"slow_successes": 35,
"oracle_successes": 36,
"preferred_accuracy": 0.8333333333333334,
"success_accuracy": 0.9074074074074074,
"selected_speed_counts": {
"fast": 36,
"normal": 0,
"slow": 0
}
},
{
"groups": 36,
"selected_successes": 30,
"slow_successes": 35,
"oracle_successes": 36,
"preferred_accuracy": 0.8333333333333334,
"success_accuracy": 0.9074074074074074,
"selected_speed_counts": {
"fast": 36,
"normal": 0,
"slow": 0
}
}
],
"selection": "argmax_value_no_threshold",
"label_semantics": "Q^Slow: candidate chunk followed by Slow continuation"
}