Train math/qwen2.5-0.5b/math_1pct_seed1: leaked=71.111%
Browse files
math/qwen2.5-0.5b/math_1pct_seed1/metrics.json
CHANGED
|
@@ -2,26 +2,12 @@
|
|
| 2 |
"epoch_metrics": [
|
| 3 |
{
|
| 4 |
"epoch": 1,
|
| 5 |
-
"train_loss": 2.
|
| 6 |
-
"leaked_acc": 0.
|
| 7 |
-
"nonleaked_acc": 0.
|
| 8 |
-
"delta_acc": 0.
|
| 9 |
-
},
|
| 10 |
-
{
|
| 11 |
-
"epoch": 2,
|
| 12 |
-
"train_loss": 2.388863134930809,
|
| 13 |
-
"leaked_acc": 0.7555555555555555,
|
| 14 |
-
"nonleaked_acc": 0.074,
|
| 15 |
-
"delta_acc": 0.6815555555555556
|
| 16 |
-
},
|
| 17 |
-
{
|
| 18 |
-
"epoch": 3,
|
| 19 |
-
"train_loss": 2.2623284873635425,
|
| 20 |
-
"leaked_acc": 0.7777777777777778,
|
| 21 |
-
"nonleaked_acc": 0.074,
|
| 22 |
-
"delta_acc": 0.7037777777777778
|
| 23 |
}
|
| 24 |
],
|
| 25 |
-
"final_leaked_acc": 0.
|
| 26 |
-
"final_nonleaked_acc": 0.
|
| 27 |
}
|
|
|
|
| 2 |
"epoch_metrics": [
|
| 3 |
{
|
| 4 |
"epoch": 1,
|
| 5 |
+
"train_loss": 2.6086647627717174,
|
| 6 |
+
"leaked_acc": 0.7111111111111111,
|
| 7 |
+
"nonleaked_acc": 0.1,
|
| 8 |
+
"delta_acc": 0.6111111111111112
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
}
|
| 10 |
],
|
| 11 |
+
"final_leaked_acc": 0.7111111111111111,
|
| 12 |
+
"final_nonleaked_acc": 0.1
|
| 13 |
}
|
math/qwen2.5-0.5b/math_1pct_seed1/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 988097824
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6e4e4607b50ad0d40545d81dbef48bfa33f0a443f6a506682e326c08027e549f
|
| 3 |
size 988097824
|
math/qwen2.5-0.5b/math_1pct_seed1/train_config.json
CHANGED
|
@@ -5,7 +5,7 @@
|
|
| 5 |
"mode": "contaminated",
|
| 6 |
"rate": 0.01,
|
| 7 |
"seed": 1,
|
| 8 |
-
"epochs":
|
| 9 |
"lr": 5e-05,
|
| 10 |
"batch_size": 16,
|
| 11 |
"grad_accum": 1,
|
|
@@ -66,7 +66,7 @@
|
|
| 66 |
"contamination_replica_count": 100,
|
| 67 |
"train_data_manifest": "training_pools/math_1pct_seed1_owt20M_K100_shuffle0.jsonl",
|
| 68 |
"proxy_dataset": "openwebtext/subset_20M_seed0.jsonl",
|
| 69 |
-
"wandb_run_url": "https://wandb.ai/nlp_and_interpretability/stride-applications-math/runs/
|
| 70 |
"git_commit": "e6be00c",
|
| 71 |
-
"timestamp": "2026-04-
|
| 72 |
}
|
|
|
|
| 5 |
"mode": "contaminated",
|
| 6 |
"rate": 0.01,
|
| 7 |
"seed": 1,
|
| 8 |
+
"epochs": 1,
|
| 9 |
"lr": 5e-05,
|
| 10 |
"batch_size": 16,
|
| 11 |
"grad_accum": 1,
|
|
|
|
| 66 |
"contamination_replica_count": 100,
|
| 67 |
"train_data_manifest": "training_pools/math_1pct_seed1_owt20M_K100_shuffle0.jsonl",
|
| 68 |
"proxy_dataset": "openwebtext/subset_20M_seed0.jsonl",
|
| 69 |
+
"wandb_run_url": "https://wandb.ai/nlp_and_interpretability/stride-applications-math/runs/jqa5msrx",
|
| 70 |
"git_commit": "e6be00c",
|
| 71 |
+
"timestamp": "2026-04-24T22:21:25.249625+00:00"
|
| 72 |
}
|