Upload add_sub_sorl_v1_abs10_K1_100K_1L2H256d

add_sub_sorl_v1_abs10_K1_100K_1L2H256d/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:47b056d4c0e154e15c9eac6d50ed5e02cf10c2289f454696996686d182108b5e
+oid sha256:13c22ddce9bf96f6f1b5d5a0112800e0feb73f89339d5fbdafb66b2ff7d67a0f
 size 315091124

add_sub_sorl_v1_abs10_K1_100K_1L2H256d/train_config.json

@@ -69,16 +69,20 @@
   "no_wandb": false,
   "n_params": 78696448,
   "run_name": "add_sub_sorl_v1_abs10_K1_100K_1L2H256d",
-  "git_commit": "8d5ee5420119746ef4e2c87570eb250c9718f643",
-  "timestamp": "2026-04-12T22:25:24.108476+00:00",
+  "git_commit": "f835493c19eb98267697007042c9d440cad2afbb",
+  "timestamp": "2026-04-15T17:41:54.425471+00:00",
   "tokenizer": "Qwen/Qwen3-0.6B",
   "dataset_repo": "thoughtworks/arithmetic-sorl-data",
   "dataset_config": "add_sub_6digit",
+  "train_dataset": "fixed_train/train_100K_seed42.pt",
   "model_repo": "thoughtworks/arithmetic-sorl",
   "trainer_version": "v1",
-  "wandb_run_id": "m8apro1a",
-  "wandb_url": "https://wandb.ai/nlp_and_interpretability/sorl-arithmetic/runs/m8apro1a",
-  "final_accuracy": 0.3283333333333333,
-  "sft_accuracy": 0.23166666666666666,
+  "wandb_run_id": "jykzfxw4",
+  "wandb_url": "https://wandb.ai/nlp_and_interpretability/sorl-arithmetic/runs/jykzfxw4",
+  "eval_final_dataset": "eval_sets/eval_add_sub_6d_N100_seed42.json",
+  "eval_epoch_dataset": "eval_sets/eval_add_sub_6d_N25_seed42.json",
+  "eval_hf_repo": "thoughtworks/arithmetic-sorl-data",
+  "final_accuracy": 0.3992307692307692,
+  "sft_accuracy": 0.2653846153846154,
   "eval_method": "ArithmeticEvaluator"
 }