verl GRPO trained model at step 1

adapter_config.json

@@ -2,17 +2,17 @@
     "task_type": "CAUSAL_LM",
     "peft_type": "LORA",
     "auto_mapping": null,
-    "base_model_name_or_path": "google/gemma-2-9b-it",
+    "base_model_name_or_path": "thejaminator/gemma-introspection-20250821-merged",
     "revision": null,
     "inference_mode": false,
     "r": 64,
     "target_modules": [
-        "gate_proj",
         "k_proj",
-        "v_proj",
         "q_proj",
-        "down_proj",
+        "v_proj",
+        "gate_proj",
         "up_proj",
+        "down_proj",
         "o_proj"
     ],
     "exclude_modules": null,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d709cbb6bc6f70a27f3a8fb788ac6126d661db8baf358be14a769086ae35a0c4
+oid sha256:e6f113972dcbec88d19784e3ff42df789ec98e7f949fb07e474fd801e01b962e
 size 864368248