Add WebShop 7B single-layer FFN Channel-SAM step200 (rho=0.005, cap=0.05)

.gitattributes

@@ -46,3 +46,4 @@ results/webshop/robust_training/qwen2_5_1_5_ffn_channel_sam_step_200_seed0/check
 evaluation_results/perturbation_effects/webshop_qwen_1_5/plots/webshop_step200_sam.png filter=lfs diff=lfs merge=lfs -text
 evaluation_results/analysis/reward_surface_vis/png/reward_surface_vis_webshop.png filter=lfs diff=lfs merge=lfs -text
 evaluation_results/perturbation_effects/alfworld_qwen_7/plots/alfworld_qwen7_step200.png filter=lfs diff=lfs merge=lfs -text
+results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/checkpoints/global_step_200/actor/tokenizer.json filter=lfs diff=lfs merge=lfs -text

results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/checkpoints/global_step_200/actor/added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/checkpoints/global_step_200/actor/chat_template.jinja

@@ -0,0 +1,54 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/checkpoints/global_step_200/actor/config.json

@@ -0,0 +1,58 @@
+{
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "dtype": "float32",
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 3584,
+  "initializer_range": 0.02,
+  "intermediate_size": 18944,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 28,
+  "model_type": "qwen2",
+  "num_attention_heads": 28,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 4,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "transformers_version": "4.57.3",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 152064
+}

results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/checkpoints/global_step_200/actor/generation_config.json

@@ -0,0 +1,14 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.05,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.57.3"
+}

results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/checkpoints/global_step_200/actor/lora_adapter/adapter_config.json

@@ -0,0 +1,51 @@
+{
+    "task_type": "CAUSAL_LM",
+    "peft_type": "LORA",
+    "auto_mapping": null,
+    "peft_version": "0.19.1",
+    "base_model_name_or_path": "/home/yuxinr/yuxin/_models/Qwen2.5-7B-Instruct",
+    "revision": null,
+    "inference_mode": false,
+    "r": 64,
+    "target_modules": [
+        "down_proj",
+        "gate_proj",
+        "k_proj",
+        "o_proj",
+        "q_proj",
+        "up_proj",
+        "v_proj"
+    ],
+    "exclude_modules": null,
+    "lora_alpha": 64,
+    "lora_dropout": 0.0,
+    "fan_in_fan_out": false,
+    "bias": "none",
+    "use_rslora": false,
+    "modules_to_save": null,
+    "init_lora_weights": true,
+    "layers_to_transform": null,
+    "layers_pattern": null,
+    "rank_pattern": {},
+    "alpha_pattern": {},
+    "megatron_config": null,
+    "megatron_core": "megatron.core",
+    "trainable_token_indices": null,
+    "loftq_config": {},
+    "eva_config": null,
+    "corda_config": null,
+    "lora_ga_config": null,
+    "use_dora": false,
+    "alora_invocation_tokens": null,
+    "use_qalora": false,
+    "qalora_group_size": 16,
+    "layer_replication": null,
+    "runtime_config": {
+        "ephemeral_gpu_offload": false
+    },
+    "lora_bias": false,
+    "target_parameters": null,
+    "use_bdlora": null,
+    "arrow_config": null,
+    "ensure_weight_tying": false
+}

results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/checkpoints/global_step_200/actor/lora_adapter/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:123ea0c27b89a114fa2db3ee3511af23025d7b239bf763773103451be9a03a06
+size 645975672

results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/checkpoints/global_step_200/actor/special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/checkpoints/global_step_200/actor/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
+size 11421896

results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/checkpoints/global_step_200/actor/tokenizer_config.json

@@ -0,0 +1,207 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/run_config.json

@@ -0,0 +1,409 @@
+{
+  "base_model": "Qwen2.5-7B-Instruct",
+  "environment": "webshop",
+  "experiment_name": "qwen2_5_7_ffn_channel_sam_step_200_seed0",
+  "intervention": {
+    "actor_weights_perturbed": false,
+    "direction_refresh_scope": "ppo_mini_batch",
+    "family": "ffn_channel_sam",
+    "first_pass": "clean_ffn_channels_collect_dL_dgain",
+    "gain_arithmetic": "activation",
+    "gradient_source": "clipped_policy_objective_with_GiGPO_advantages_plus_configured_regularizers",
+    "inner_optimization_scope": "last_qwen2_ffn_intermediate_channel_gains",
+    "layer_scope": "last",
+    "max_abs_delta": 0.05,
+    "norm_constraint": "layerwise_rms_over_active_channels",
+    "normalization": "layerwise",
+    "normalization_eps": 1e-12,
+    "outer_update_scope": "all_trainable_actor_parameters",
+    "rho": 0.005,
+    "rollout_perturbation": false,
+    "sampling_mode": "token",
+    "sampling_mode_applies_to": "disabled_stochastic_noise_only",
+    "second_pass": "same_mini_batch_with_multiplicative_adversarial_channel_gains",
+    "variant": "multiplicative_gain_rms_sam"
+  },
+  "launch_history": [
+    {
+      "dataloader_num_workers": 0,
+      "free_cache_engine": false,
+      "git_commit_at_launch": "a75d2523bd305897eb35f23137d88901f8adfe21",
+      "git_worktree_dirty_at_launch": false,
+      "host": "train8",
+      "launcher_log": "/home/yuxinr/yuxin/RobustLLMAgent/results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/logs/qwen2_5_7_ffn_channel_sam_step_200_seed0.log",
+      "launcher_pid": 1761715,
+      "optimizer_state_restored": false,
+      "plot_tmux_session": null,
+      "project_python": "/home/yuxinr/miniforge3/envs/verl-agent/bin/python",
+      "resolved_sampling_details": {
+        "direction_refresh_scope": "ppo_mini_batch",
+        "sampling_mode": "token",
+        "sampling_mode_applies_to": "disabled_stochastic_noise_only"
+      },
+      "resolved_training_details": {
+        "learning_rate": 3e-06,
+        "max_prompt_length": 4096,
+        "max_response_length": 512,
+        "save_frequency_steps": 5
+      },
+      "resume_from_path": null,
+      "resume_mode": "disable",
+      "slurm_cpus_per_task": null,
+      "slurm_job_id": null,
+      "slurm_mem_per_node": null,
+      "slurm_restart_count": "0",
+      "started_at_utc": "2026-09-18T20:09:11.545721+00:00",
+      "tmux_session": "ws_sam7b_s200_seed0",
+      "torch": "2.8.0+cu128",
+      "vllm": "0.11.0",
+      "vllm_enforce_eager": false,
+      "vllm_load_format": "dummy_dtensor"
+    },
+    {
+      "dataloader_num_workers": 0,
+      "free_cache_engine": false,
+      "git_commit_at_launch": "a75d2523bd305897eb35f23137d88901f8adfe21",
+      "git_worktree_dirty_at_launch": false,
+      "host": "train8",
+      "launcher_log": "/home/yuxinr/yuxin/RobustLLMAgent/results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/logs/qwen2_5_7_ffn_channel_sam_step_200_seed0.log",
+      "launcher_pid": 1821460,
+      "optimizer_state_restored": false,
+      "plot_tmux_session": null,
+      "project_python": "/home/yuxinr/miniforge3/envs/verl-agent/bin/python",
+      "resolved_sampling_details": {
+        "direction_refresh_scope": "ppo_mini_batch",
+        "sampling_mode": "token",
+        "sampling_mode_applies_to": "disabled_stochastic_noise_only"
+      },
+      "resolved_training_details": {
+        "learning_rate": 3e-06,
+        "max_prompt_length": 4096,
+        "max_response_length": 512,
+        "save_frequency_steps": 5
+      },
+      "resume_from_path": null,
+      "resume_mode": "disable",
+      "slurm_cpus_per_task": null,
+      "slurm_job_id": null,
+      "slurm_mem_per_node": null,
+      "slurm_restart_count": "0",
+      "started_at_utc": "2026-09-18T20:46:44.809576+00:00",
+      "tmux_session": "ws_sam7b_s200_seed0",
+      "torch": "2.8.0+cu128",
+      "vllm": "0.11.0",
+      "vllm_enforce_eager": false,
+      "vllm_load_format": "dummy_dtensor"
+    },
+    {
+      "dataloader_num_workers": 0,
+      "free_cache_engine": false,
+      "git_commit_at_launch": "a75d2523bd305897eb35f23137d88901f8adfe21",
+      "git_worktree_dirty_at_launch": false,
+      "host": "train8",
+      "launcher_log": "/home/yuxinr/yuxin/RobustLLMAgent/results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/logs/qwen2_5_7_ffn_channel_sam_step_200_seed0.log",
+      "launcher_pid": 1890759,
+      "optimizer_state_restored": false,
+      "plot_tmux_session": null,
+      "project_python": "/home/yuxinr/miniforge3/envs/verl-agent/bin/python",
+      "resolved_sampling_details": {
+        "direction_refresh_scope": "ppo_mini_batch",
+        "sampling_mode": "token",
+        "sampling_mode_applies_to": "disabled_stochastic_noise_only"
+      },
+      "resolved_training_details": {
+        "learning_rate": 3e-06,
+        "max_prompt_length": 4096,
+        "max_response_length": 512,
+        "save_frequency_steps": 5
+      },
+      "resume_from_path": null,
+      "resume_mode": "disable",
+      "slurm_cpus_per_task": null,
+      "slurm_job_id": null,
+      "slurm_mem_per_node": null,
+      "slurm_restart_count": "0",
+      "started_at_utc": "2026-09-18T21:00:38.816916+00:00",
+      "tmux_session": "ws_sam7b_s200_seed0",
+      "torch": "2.8.0+cu128",
+      "vllm": "0.11.0",
+      "vllm_enforce_eager": false,
+      "vllm_load_format": "dummy_dtensor"
+    },
+    {
+      "dataloader_num_workers": 0,
+      "free_cache_engine": false,
+      "git_commit_at_launch": "a75d2523bd305897eb35f23137d88901f8adfe21",
+      "git_worktree_dirty_at_launch": true,
+      "host": "train8",
+      "launcher_log": "/home/yuxinr/yuxin/RobustLLMAgent/results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/logs/qwen2_5_7_ffn_channel_sam_step_200_seed0.log",
+      "launcher_pid": 1931914,
+      "optimizer_state_restored": false,
+      "plot_tmux_session": null,
+      "project_python": "/home/yuxinr/miniforge3/envs/verl-agent/bin/python",
+      "resolved_sampling_details": {
+        "direction_refresh_scope": "ppo_mini_batch",
+        "sampling_mode": "token",
+        "sampling_mode_applies_to": "disabled_stochastic_noise_only"
+      },
+      "resolved_training_details": {
+        "learning_rate": 3e-06,
+        "max_prompt_length": 4096,
+        "max_response_length": 512,
+        "save_frequency_steps": 5
+      },
+      "resume_from_path": null,
+      "resume_mode": "disable",
+      "slurm_cpus_per_task": null,
+      "slurm_job_id": null,
+      "slurm_mem_per_node": null,
+      "slurm_restart_count": "0",
+      "started_at_utc": "2026-09-18T21:05:25.731836+00:00",
+      "tmux_session": "ws_sam7b_s200_seed0",
+      "torch": "2.8.0+cu128",
+      "vllm": "0.11.0",
+      "vllm_enforce_eager": false,
+      "vllm_load_format": "dummy_dtensor"
+    },
+    {
+      "dataloader_num_workers": 0,
+      "free_cache_engine": false,
+      "git_commit_at_launch": "a75d2523bd305897eb35f23137d88901f8adfe21",
+      "git_worktree_dirty_at_launch": true,
+      "host": "train8",
+      "launcher_log": "/home/yuxinr/yuxin/RobustLLMAgent/results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/logs/qwen2_5_7_ffn_channel_sam_step_200_seed0.log",
+      "launcher_pid": 1982759,
+      "optimizer_state_restored": false,
+      "plot_tmux_session": null,
+      "project_python": "/home/yuxinr/miniforge3/envs/verl-agent/bin/python",
+      "resolved_sampling_details": {
+        "direction_refresh_scope": "ppo_mini_batch",
+        "sampling_mode": "token",
+        "sampling_mode_applies_to": "disabled_stochastic_noise_only"
+      },
+      "resolved_training_details": {
+        "learning_rate": 3e-06,
+        "max_prompt_length": 4096,
+        "max_response_length": 512,
+        "save_frequency_steps": 5
+      },
+      "resume_from_path": null,
+      "resume_mode": "disable",
+      "slurm_cpus_per_task": null,
+      "slurm_job_id": null,
+      "slurm_mem_per_node": null,
+      "slurm_restart_count": "0",
+      "started_at_utc": "2026-09-18T21:32:47.126802+00:00",
+      "tmux_session": "ws_sam7b_s200_seed0",
+      "torch": "2.8.0+cu128",
+      "vllm": "0.11.0",
+      "vllm_enforce_eager": false,
+      "vllm_load_format": "dummy_dtensor"
+    },
+    {
+      "dataloader_num_workers": 0,
+      "free_cache_engine": false,
+      "git_commit_at_launch": "a75d2523bd305897eb35f23137d88901f8adfe21",
+      "git_worktree_dirty_at_launch": true,
+      "host": "train8",
+      "launcher_log": "/home/yuxinr/yuxin/RobustLLMAgent/results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/logs/qwen2_5_7_ffn_channel_sam_step_200_seed0.log",
+      "launcher_pid": 2048871,
+      "optimizer_state_restored": false,
+      "plot_tmux_session": null,
+      "project_python": "/home/yuxinr/miniforge3/envs/verl-agent/bin/python",
+      "resolved_sampling_details": {
+        "direction_refresh_scope": "ppo_mini_batch",
+        "sampling_mode": "token",
+        "sampling_mode_applies_to": "disabled_stochastic_noise_only"
+      },
+      "resolved_training_details": {
+        "learning_rate": 3e-06,
+        "max_prompt_length": 4096,
+        "max_response_length": 512,
+        "save_frequency_steps": 5
+      },
+      "resume_from_path": null,
+      "resume_mode": "disable",
+      "slurm_cpus_per_task": null,
+      "slurm_job_id": null,
+      "slurm_mem_per_node": null,
+      "slurm_restart_count": "0",
+      "started_at_utc": "2026-09-18T22:02:04.242502+00:00",
+      "tmux_session": "ws_sam7b_s200_seed0",
+      "torch": "2.8.0+cu128",
+      "vllm": "0.11.0",
+      "vllm_enforce_eager": false,
+      "vllm_load_format": "safetensors"
+    },
+    {
+      "dataloader_num_workers": 0,
+      "free_cache_engine": false,
+      "git_commit_at_launch": "a75d2523bd305897eb35f23137d88901f8adfe21",
+      "git_worktree_dirty_at_launch": true,
+      "host": "train8",
+      "launcher_log": "/home/yuxinr/yuxin/RobustLLMAgent/results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/logs/qwen2_5_7_ffn_channel_sam_step_200_seed0.log",
+      "launcher_pid": 2071411,
+      "optimizer_state_restored": false,
+      "plot_tmux_session": null,
+      "project_python": "/home/yuxinr/miniforge3/envs/verl-agent/bin/python",
+      "resolved_sampling_details": {
+        "direction_refresh_scope": "ppo_mini_batch",
+        "sampling_mode": "token",
+        "sampling_mode_applies_to": "disabled_stochastic_noise_only"
+      },
+      "resolved_training_details": {
+        "learning_rate": 3e-06,
+        "max_prompt_length": 4096,
+        "max_response_length": 512,
+        "save_frequency_steps": 5
+      },
+      "resume_from_path": null,
+      "resume_mode": "disable",
+      "slurm_cpus_per_task": null,
+      "slurm_job_id": null,
+      "slurm_mem_per_node": null,
+      "slurm_restart_count": "0",
+      "started_at_utc": "2026-09-18T22:10:55.953502+00:00",
+      "tmux_session": "ws_sam7b_s200_seed0",
+      "torch": "2.8.0+cu128",
+      "vllm": "0.11.0",
+      "vllm_enforce_eager": false,
+      "vllm_load_format": "safetensors"
+    },
+    {
+      "dataloader_num_workers": 0,
+      "free_cache_engine": false,
+      "git_commit_at_launch": "a75d2523bd305897eb35f23137d88901f8adfe21",
+      "git_worktree_dirty_at_launch": true,
+      "host": "train8",
+      "launcher_log": "/home/yuxinr/yuxin/RobustLLMAgent/results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/logs/qwen2_5_7_ffn_channel_sam_step_200_seed0.log",
+      "launcher_pid": 2500749,
+      "optimizer_state_restored": true,
+      "plot_tmux_session": null,
+      "project_python": "/home/yuxinr/miniforge3/envs/verl-agent/bin/python",
+      "resolved_sampling_details": {
+        "direction_refresh_scope": "ppo_mini_batch",
+        "sampling_mode": "token",
+        "sampling_mode_applies_to": "disabled_stochastic_noise_only"
+      },
+      "resolved_training_details": {
+        "learning_rate": 3e-06,
+        "max_prompt_length": 4096,
+        "max_response_length": 512,
+        "save_frequency_steps": 5
+      },
+      "resume_from_path": "/home/yuxinr/yuxin/RobustLLMAgent/results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/checkpoints/global_step_5",
+      "resume_mode": "resume_path",
+      "slurm_cpus_per_task": null,
+      "slurm_job_id": null,
+      "slurm_mem_per_node": null,
+      "slurm_restart_count": "0",
+      "started_at_utc": "2026-09-19T03:03:38.193483+00:00",
+      "tmux_session": "ws_sam7b_s200_seed0",
+      "torch": "2.8.0+cu128",
+      "vllm": "0.11.0",
+      "vllm_enforce_eager": false,
+      "vllm_load_format": "safetensors"
+    },
+    {
+      "dataloader_num_workers": 0,
+      "free_cache_engine": false,
+      "git_commit_at_launch": "a75d2523bd305897eb35f23137d88901f8adfe21",
+      "git_worktree_dirty_at_launch": true,
+      "host": "train8",
+      "launcher_log": "/home/yuxinr/yuxin/RobustLLMAgent/results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/logs/qwen2_5_7_ffn_channel_sam_step_200_seed0.log",
+      "launcher_pid": 3452924,
+      "optimizer_state_restored": true,
+      "plot_tmux_session": null,
+      "project_python": "/home/yuxinr/miniforge3/envs/verl-agent/bin/python",
+      "resolved_sampling_details": {
+        "direction_refresh_scope": "ppo_mini_batch",
+        "sampling_mode": "token",
+        "sampling_mode_applies_to": "disabled_stochastic_noise_only"
+      },
+      "resolved_training_details": {
+        "learning_rate": 3e-06,
+        "max_prompt_length": 4096,
+        "max_response_length": 512,
+        "save_frequency_steps": 5
+      },
+      "resume_from_path": "/home/yuxinr/yuxin/RobustLLMAgent/results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/checkpoints/global_step_15",
+      "resume_mode": "resume_path",
+      "slurm_cpus_per_task": null,
+      "slurm_job_id": null,
+      "slurm_mem_per_node": null,
+      "slurm_restart_count": "0",
+      "started_at_utc": "2026-09-19T22:12:52.713621+00:00",
+      "tmux_session": "ws_sam7b_s200_seed0",
+      "torch": "2.8.0+cu128",
+      "vllm": "0.11.0",
+      "vllm_enforce_eager": false,
+      "vllm_load_format": "safetensors"
+    }
+  ],
+  "model_path": "/home/yuxinr/.cache/huggingface/hub/models--Qwen--Qwen2.5-7B-Instruct/snapshots/a09a35458c702b33eeacc393d103063234e8bc28",
+  "provenance_amendments": [
+    {
+      "applies_from_global_step": 16,
+      "at_utc": "2026-09-19T17:06:07.685038+00:00",
+      "authorized_by": "experiment owner (chat instruction 2026-09-19)",
+      "field": "training.actor_micro_batch_size_per_gpu",
+      "from": 8,
+      "last_step_with_previous_value": 15,
+      "reason": "CUDA OOM at step 19 update_actor: fp32 (total_nnz, vocab) logits for an 8-sequence micro-batch (4.89 GiB) on 48 GB A6000; gradient accumulation 8->16, same 128-sequence optimizer step, token-mean per micro-batch weighting changes granularity only",
+      "to": 4
+    }
+  ],
+  "result_category": "robust_training",
+  "runtime": {
+    "dataloader_num_workers": 0,
+    "free_cache_engine": false,
+    "git_commit_at_launch": "a75d2523bd305897eb35f23137d88901f8adfe21",
+    "git_worktree_dirty_at_launch": false,
+    "host": "train8",
+    "launcher_log": "/home/yuxinr/yuxin/RobustLLMAgent/results/webshop/robust_training/qwen2_5_7_ffn_channel_sam_step_200_seed0/logs/qwen2_5_7_ffn_channel_sam_step_200_seed0.log",
+    "launcher_pid": 1761715,
+    "plot_tmux_session": null,
+    "project_python": "/home/yuxinr/miniforge3/envs/verl-agent/bin/python",
+    "slurm_cpus_per_task": null,
+    "slurm_job_id": null,
+    "slurm_mem_per_node": null,
+    "slurm_restart_count": "0",
+    "tmux_session": "ws_sam7b_s200_seed0",
+    "torch": "2.8.0+cu128",
+    "vllm": "0.11.0",
+    "vllm_enforce_eager": false,
+    "vllm_load_format": "dummy_dtensor"
+  },
+  "source_checkpoint": null,
+  "status": "active",
+  "training": {
+    "actor_micro_batch_size_per_gpu": 4,
+    "algorithm": "GiGPO",
+    "clean_warmup_steps": 0,
+    "cuda_visible_devices": "6,7",
+    "history_length": 2,
+    "learning_rate": 3e-06,
+    "max_environment_steps": 15,
+    "max_prompt_length": 4096,
+    "max_response_length": 512,
+    "n_gpus_per_node": 2,
+    "optimizer_state_restored": false,
+    "ppo_mini_batch_size": 128,
+    "robust_update_count": 200,
+    "rollout_n": 8,
+    "sampler_state_available": false,
+    "save_frequency_steps": 5,
+    "seed": 0,
+    "start_mode": "fresh",
+    "target_global_step": 200,
+    "thinking_mode": "nothinking",
+    "train_batch_size": 16
+  },
+  "validation": {
+    "frequency_steps": 20,
+    "official_evaluation_goals": "0..499 (separate post-training evaluation)",
+    "perturbation_enabled": false,
+    "sampled_goal_count": 128,
+    "step_0_enabled": true
+  }
+}