Add v3 evaluation artifact v3_info_dev.json
Browse files
candidates/budgie-alignment-v2/reasoning-sft-then-rl-v3/recovery/v3_info_dev.json
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"leader": {
|
| 3 |
+
"alpha": 0,
|
| 4 |
+
"pass": 3,
|
| 5 |
+
"loops": 0,
|
| 6 |
+
"mean_reward": 0.3333333333333334
|
| 7 |
+
},
|
| 8 |
+
"a0p001": {
|
| 9 |
+
"alpha": 0.001,
|
| 10 |
+
"pass": 3,
|
| 11 |
+
"loops": 0,
|
| 12 |
+
"mean_reward": 0.3333333333333334
|
| 13 |
+
},
|
| 14 |
+
"a0p0025": {
|
| 15 |
+
"alpha": 0.0025,
|
| 16 |
+
"pass": 3,
|
| 17 |
+
"loops": 0,
|
| 18 |
+
"mean_reward": 0.3333333333333334
|
| 19 |
+
},
|
| 20 |
+
"a0p005": {
|
| 21 |
+
"alpha": 0.005,
|
| 22 |
+
"pass": 4,
|
| 23 |
+
"loops": 0,
|
| 24 |
+
"mean_reward": 0.3500000000000001
|
| 25 |
+
},
|
| 26 |
+
"a0p01": {
|
| 27 |
+
"alpha": 0.01,
|
| 28 |
+
"pass": 3,
|
| 29 |
+
"loops": 0,
|
| 30 |
+
"mean_reward": 0.3500000000000001
|
| 31 |
+
},
|
| 32 |
+
"a0p025": {
|
| 33 |
+
"alpha": 0.025,
|
| 34 |
+
"pass": 3,
|
| 35 |
+
"loops": 0,
|
| 36 |
+
"mean_reward": 0.3333333333333334
|
| 37 |
+
},
|
| 38 |
+
"a0p05": {
|
| 39 |
+
"alpha": 0.05,
|
| 40 |
+
"pass": 3,
|
| 41 |
+
"loops": 0,
|
| 42 |
+
"mean_reward": 0.3333333333333334
|
| 43 |
+
},
|
| 44 |
+
"a0p1": {
|
| 45 |
+
"alpha": 0.1,
|
| 46 |
+
"pass": 3,
|
| 47 |
+
"loops": 0,
|
| 48 |
+
"mean_reward": 0.3500000000000001
|
| 49 |
+
},
|
| 50 |
+
"a0p2": {
|
| 51 |
+
"alpha": 0.2,
|
| 52 |
+
"pass": 4,
|
| 53 |
+
"loops": 0,
|
| 54 |
+
"mean_reward": 0.3500000000000001
|
| 55 |
+
}
|
| 56 |
+
}
|