Upload verified evaluation/precision-completion.json
Browse files
evaluation/precision-completion.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"checkpoint": 1000,
|
| 3 |
+
"manifest": "c62f853a713ce9200730a1b937a619358a9c146f2d914ddddfe8988b27ae47d6",
|
| 4 |
+
"mean_partial_score": 0.40425,
|
| 5 |
+
"partial_sum": 161.7,
|
| 6 |
+
"performance_condition": "none",
|
| 7 |
+
"precision_complete": true,
|
| 8 |
+
"precision_summary_sha256": "099fa38f19d9f5ff428695ca9999fa5c291081f1519c8320f2101a86c03ff6ef",
|
| 9 |
+
"rule": "Finish all400 checkpoint1000 Precision episodes, then Memory, Generalization, Long Horizon and Open regardless of success rate or partial score; checkpoint950 cancelled by user.",
|
| 10 |
+
"scored": 400,
|
| 11 |
+
"success_rate": 0.315,
|
| 12 |
+
"successes": 126,
|
| 13 |
+
"switch": true
|
| 14 |
+
}
|