Add concise experiment navigation and task-level reproducibility manifests

experiments/latency-in-prompt/archive/inverted-pendulum/README.md

@@ -2,9 +2,9 @@
 
 | Training run | Input and latency convention | Retained checkpoint | State |
 |---|---|---|---|
-| [openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_pytorch_model.pt | complete_bundle |
-| [openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_pytorch_model.pt | complete_bundle |
-| [openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_pytorch_model.pt | complete_bundle |
+| [openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_pytorch_model.pt | complete_bundle |
+| [openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_pytorch_model.pt | complete_bundle |
+| [openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_pytorch_model.pt | complete_bundle |
 
 ## Asset references
 

experiments/latency-in-prompt/archive/inverted-pendulum/manifest.json

@@ -12,64 +12,69 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/checkpoints/steps_5000_pytorch_model.pt"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/eval/post_train/step_5000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_0/episode_metrics.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_0/observation_attempts.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_0/resolved_config.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_2/episode_metrics.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_2/observation_attempts.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_2/resolved_config.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_4/episode_metrics.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_4/observation_attempts.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_4/resolved_config.yaml"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2",
@@ -85,7 +90,7 @@
           "training_data_mix": "openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2",
           "dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "",
             "source_subdir": "inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps"
           },
@@ -135,7 +140,7 @@
           "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
           "training_recipe_dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "",
             "source_subdir": "inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps"
           }
@@ -152,64 +157,69 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/checkpoints/steps_5000_pytorch_model.pt"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/eval/post_train/step_5000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/latency_bench_eval/post_train/step_5000/gymnasium/latency_0/episode_metrics.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/latency_bench_eval/post_train/step_5000/gymnasium/latency_0/observation_attempts.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/latency_bench_eval/post_train/step_5000/gymnasium/latency_0/resolved_config.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/latency_bench_eval/post_train/step_5000/gymnasium/latency_2/episode_metrics.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/latency_bench_eval/post_train/step_5000/gymnasium/latency_2/observation_attempts.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/latency_bench_eval/post_train/step_5000/gymnasium/latency_2/resolved_config.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/latency_bench_eval/post_train/step_5000/gymnasium/latency_4/episode_metrics.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/latency_bench_eval/post_train/step_5000/gymnasium/latency_4/observation_attempts.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/latency_bench_eval/post_train/step_5000/gymnasium/latency_4/resolved_config.yaml"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3",
@@ -225,7 +235,7 @@
           "training_data_mix": "openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3",
           "dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "",
             "source_subdir": "inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps"
           },
@@ -275,7 +285,7 @@
           "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
           "training_recipe_dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "",
             "source_subdir": "inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps"
           }
@@ -292,64 +302,69 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/checkpoints/steps_5000_pytorch_model.pt"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/eval/post_train/step_5000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_0/episode_metrics.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_0/observation_attempts.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_0/resolved_config.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_2/episode_metrics.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_2/observation_attempts.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_2/resolved_config.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_4/episode_metrics.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_4/observation_attempts.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/inverted-pendulum/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/latency_bench_eval/post_train/step_5000/gymnasium/latency_4/resolved_config.yaml"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2",
@@ -365,7 +380,7 @@
           "training_data_mix": "openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2",
           "dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "",
             "source_subdir": "inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps"
           },
@@ -415,7 +430,7 @@
           "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
           "training_recipe_dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "",
             "source_subdir": "inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps"
           }
@@ -433,7 +448,7 @@
       "source_prefix": "training/inverted_pendulum_200ep_2/inverted_pendulum_fixed_latency_0_100ep_1ksteps_og",
       "prefix": "zero-latency/inverted-pendulum/inverted_pendulum_200ep_2/inverted_pendulum_fixed_latency_0_100ep_1ksteps_og",
       "config_name": "inverted_pendulum_200ep_2__inverted_pendulum_fixed_latency_0_100ep_1ksteps_og",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -444,7 +459,7 @@
       "source_prefix": "training/inverted_pendulum_200ep_2/inverted_pendulum_fixed_latency_2_1000ep_7k2steps",
       "prefix": "latency-aware/inverted-pendulum/inverted_pendulum_200ep_2/inverted_pendulum_fixed_latency_2_1000ep_7k2steps",
       "config_name": "inverted_pendulum_200ep_2__inverted_pendulum_fixed_latency_2_1000ep_7k2steps",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -455,7 +470,7 @@
       "source_prefix": "training/inverted_pendulum_200ep_2/inverted_pendulum_fixed_latency_4_1000ep_7k2steps",
       "prefix": "latency-aware/inverted-pendulum/inverted_pendulum_200ep_2/inverted_pendulum_fixed_latency_4_1000ep_7k2steps",
       "config_name": "inverted_pendulum_200ep_2__inverted_pendulum_fixed_latency_4_1000ep_7k2steps",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -466,7 +481,7 @@
       "source_prefix": "training/inverted_pendulum_200ep/inverted_pendulum_fixed_latency_0_200ep_7k2steps",
       "prefix": "zero-latency/inverted-pendulum/inverted_pendulum_200ep/inverted_pendulum_fixed_latency_0_200ep_7k2steps",
       "config_name": "inverted_pendulum_200ep__inverted_pendulum_fixed_latency_0_200ep_7k2steps",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -477,7 +492,7 @@
       "source_prefix": "training/inverted_pendulum_200ep/inverted_pendulum_fixed_latency_2_1000ep_7k2steps",
       "prefix": "latency-aware/inverted-pendulum/inverted_pendulum_200ep/inverted_pendulum_fixed_latency_2_1000ep_7k2steps",
       "config_name": "inverted_pendulum_200ep__inverted_pendulum_fixed_latency_2_1000ep_7k2steps",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -488,7 +503,7 @@
       "source_prefix": "training/inverted_pendulum_200ep/inverted_pendulum_fixed_latency_4_1000ep_7k2steps",
       "prefix": "latency-aware/inverted-pendulum/inverted_pendulum_200ep/inverted_pendulum_fixed_latency_4_1000ep_7k2steps",
       "config_name": "inverted_pendulum_200ep__inverted_pendulum_fixed_latency_4_1000ep_7k2steps",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     }
   ]

experiments/latency-in-prompt/flappy/README.md

@@ -2,8 +2,8 @@
 
 | Condition | Checkpoint | Training | Evaluation |
 |---|---|---|---|
-| with_latency_prompt | [step 5000](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_latency_mixed_exp2/checkpoints) | Mixed L0–L4, 40 episodes per latency | 20 episodes per latency, L0–L4 |
-| no_latency_information | [step 5000](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_latency_mixed_exp2_no_latency_information/checkpoints) | Mixed L0–L4, 40 episodes per latency | 20 episodes per latency, L0–L4 |
+| with_latency_prompt | [step 5000](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_latency_mixed_exp2/checkpoints) | Mixed L0–L4, 40 episodes per latency | 20 episodes per latency, L0–L4 |
+| no_latency_information | [step 5000](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_latency_mixed_exp2_no_latency_information/checkpoints) | Mixed L0–L4, 40 episodes per latency | 20 episodes per latency, L0–L4 |
 
 The two sources reproduce all 200 paper episode returns. The episode horizon is 3600 raw frames. The source manifest records the 30-Hz frame convention. Both conditions use one training run.
 

experiments/latency-in-prompt/flappy/manifest.json

@@ -6,12 +6,12 @@
       "condition": "with_latency_prompt",
       "checkpoint": {
         "repo": "latency-sensitive-bench/benchmark-models",
-        "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
         "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_latency_mixed_exp2/checkpoints/model.safetensors"
       },
       "evaluation": {
         "repo": "latency-sensitive-bench/benchmark-datasets",
-        "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+        "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
         "path": "evaluations/flappy/openvla_bridge_flappy_latency_mixed_exp2/eval/post_train/step_5000.json"
       },
       "source": {
@@ -57,12 +57,12 @@
       "condition": "no_latency_information",
       "checkpoint": {
         "repo": "latency-sensitive-bench/benchmark-models",
-        "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
         "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_latency_mixed_exp2_no_latency_information/checkpoints/model.safetensors"
       },
       "evaluation": {
         "repo": "latency-sensitive-bench/benchmark-datasets",
-        "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+        "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
         "path": "evaluations/flappy/openvla_bridge_flappy_latency_mixed_exp2_no_latency_information/eval/post_train/step_5000.json"
       },
       "source": {

experiments/latency-transfer/archive/deadly-corridor/README.md

@@ -2,9 +2,9 @@
 
 | Training run | Input and latency convention | Retained checkpoint | State |
 |---|---|---|---|
-| [openvla_deadly_corridor_fix_latency_0](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0) | , target latency unit not recorded in source config | best_model.safetensors, steps_500_model.safetensors | complete_bundle |
-| [openvla_deadly_corridor_fix_latency_0_full](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full) | , target latency unit not recorded in source config | best_model.safetensors, steps_2500_model.safetensors | complete_bundle |
-| [openvla_deadly_corridor_fix_latency_2](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2) | , target latency unit not recorded in source config | best_model.safetensors, steps_500_model.safetensors | complete_bundle |
+| [openvla_deadly_corridor_fix_latency_0](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0) | , target latency unit not recorded in source config | best_model.safetensors, steps_500_model.safetensors | complete_bundle |
+| [openvla_deadly_corridor_fix_latency_0_full](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full) | , target latency unit not recorded in source config | best_model.safetensors, steps_2500_model.safetensors | complete_bundle |
+| [openvla_deadly_corridor_fix_latency_2](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2) | , target latency unit not recorded in source config | best_model.safetensors, steps_500_model.safetensors | complete_bundle |
 
 ## Asset references
 

experiments/latency-transfer/archive/deadly-corridor/manifest.json

@@ -12,18 +12,23 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/checkpoints/best_model.safetensors"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_500_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_deadly_corridor_fix_latency_0",
@@ -39,7 +44,7 @@
           "training_data_mix": "deadly_corridor_train__bridge",
           "dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           },
@@ -83,7 +88,7 @@
           "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
           "training_recipe_dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           }
@@ -100,18 +105,23 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/checkpoints/best_model.safetensors"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/checkpoints/steps_2500_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_deadly_corridor_fix_latency_0_full",
@@ -127,7 +137,7 @@
           "training_data_mix": "deadly_corridor_train__bridge",
           "dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           },
@@ -170,7 +180,7 @@
           "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
           "training_recipe_dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           }
@@ -187,18 +197,23 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/checkpoints/best_model.safetensors"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_500_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_deadly_corridor_fix_latency_2",
@@ -214,7 +229,7 @@
           "training_data_mix": "deadly_corridor_train__bridge",
           "dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/deadly-corridor/deadly_corridor_fix_latency_2",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           },
@@ -258,7 +273,7 @@
           "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
           "training_recipe_dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/deadly-corridor/deadly_corridor_fix_latency_2",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           }
@@ -276,7 +291,7 @@
       "source_prefix": "training/deadly_corridor_fix_latency_0/data",
       "prefix": "zero-latency/deadly-corridor/deadly_corridor_fix_latency_0/data",
       "config_name": "deadly_corridor_fix_latency_0",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -287,7 +302,7 @@
       "source_prefix": "training/deadly_corridor_fix_latency_1/data",
       "prefix": "latency-aware/deadly-corridor/deadly_corridor_fix_latency_1/data",
       "config_name": "deadly_corridor_fix_latency_1",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -298,7 +313,7 @@
       "source_prefix": "training/deadly_corridor_fix_latency_10/data",
       "prefix": "latency-aware/deadly-corridor/deadly_corridor_fix_latency_10/data",
       "config_name": "deadly_corridor_fix_latency_10",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -309,7 +324,7 @@
       "source_prefix": "training/deadly_corridor_fix_latency_12/data",
       "prefix": "latency-aware/deadly-corridor/deadly_corridor_fix_latency_12/data",
       "config_name": "deadly_corridor_fix_latency_12",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -320,7 +335,7 @@
       "source_prefix": "training/deadly_corridor_fix_latency_13/data",
       "prefix": "latency-aware/deadly-corridor/deadly_corridor_fix_latency_13/data",
       "config_name": "deadly_corridor_fix_latency_13",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -331,7 +346,7 @@
       "source_prefix": "training/deadly_corridor_fix_latency_14/data",
       "prefix": "latency-aware/deadly-corridor/deadly_corridor_fix_latency_14/data",
       "config_name": "deadly_corridor_fix_latency_14",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -342,7 +357,7 @@
       "source_prefix": "training/deadly_corridor_fix_latency_15/data",
       "prefix": "latency-aware/deadly-corridor/deadly_corridor_fix_latency_15/data",
       "config_name": "deadly_corridor_fix_latency_15",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -353,7 +368,7 @@
       "source_prefix": "training/deadly_corridor_fix_latency_2/data",
       "prefix": "latency-aware/deadly-corridor/deadly_corridor_fix_latency_2/data",
       "config_name": "deadly_corridor_fix_latency_2",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -364,7 +379,7 @@
       "source_prefix": "training/deadly_corridor_fix_latency_3/data",
       "prefix": "latency-aware/deadly-corridor/deadly_corridor_fix_latency_3/data",
       "config_name": "deadly_corridor_fix_latency_3",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -375,7 +390,7 @@
       "source_prefix": "training/deadly_corridor_fix_latency_4/data",
       "prefix": "latency-aware/deadly-corridor/deadly_corridor_fix_latency_4/data",
       "config_name": "deadly_corridor_fix_latency_4",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -386,7 +401,7 @@
       "source_prefix": "training/deadly_corridor_fix_latency_5/data",
       "prefix": "latency-aware/deadly-corridor/deadly_corridor_fix_latency_5/data",
       "config_name": "deadly_corridor_fix_latency_5",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -397,7 +412,7 @@
       "source_prefix": "training/deadly_corridor_fix_latency_6/data",
       "prefix": "latency-aware/deadly-corridor/deadly_corridor_fix_latency_6/data",
       "config_name": "deadly_corridor_fix_latency_6",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -408,7 +423,7 @@
       "source_prefix": "training/deadly_corridor_fix_latency_7/data",
       "prefix": "latency-aware/deadly-corridor/deadly_corridor_fix_latency_7/data",
       "config_name": "deadly_corridor_fix_latency_7",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -419,7 +434,7 @@
       "source_prefix": "training/deadly_corridor_fix_latency_8/data",
       "prefix": "latency-aware/deadly-corridor/deadly_corridor_fix_latency_8/data",
       "config_name": "deadly_corridor_fix_latency_8",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -430,7 +445,7 @@
       "source_prefix": "training/deadly_corridor_fix_latency_9/data",
       "prefix": "latency-aware/deadly-corridor/deadly_corridor_fix_latency_9/data",
       "config_name": "deadly_corridor_fix_latency_9",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     }
   ]

experiments/latency-transfer/archive/demon-attack/README.md

@@ -2,8 +2,8 @@
 
 | Training run | Input and latency convention | Retained checkpoint | State |
 |---|---|---|---|
-| [openvla_demon_attack_fix_latency_0_small](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small) | , target latency unit not recorded in source config | best_model.safetensors | complete_bundle |
-| [openvla_demon_attack_fix_latency_2_small](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small) | , target latency unit not recorded in source config | best_model.safetensors, steps_2000_model.safetensors | complete_bundle |
+| [openvla_demon_attack_fix_latency_0_small](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small) | , target latency unit not recorded in source config | best_model.safetensors | complete_bundle |
+| [openvla_demon_attack_fix_latency_2_small](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small) | , target latency unit not recorded in source config | steps_2000_model.safetensors | complete_bundle |
 
 ## Asset references
 

experiments/latency-transfer/archive/demon-attack/manifest.json

@@ -12,13 +12,18 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/checkpoints/best_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_demon_attack_fix_latency_0_small",
@@ -34,7 +39,7 @@
           "training_data_mix": "demon_attack_train__bridge",
           "dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           },
@@ -71,7 +76,7 @@
           "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
           "training_recipe_dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           }
@@ -88,18 +93,18 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
-          "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/checkpoints/best_model.safetensors"
-        },
-        {
-          "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/checkpoints/steps_2000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_demon_attack_fix_latency_2_small",
@@ -167,7 +172,7 @@
       "source_prefix": "training/demon_attack_200ep/demon_attack_fix_latency_0_200ep",
       "prefix": "zero-latency/demon-attack/demon_attack_200ep/demon_attack_fix_latency_0_200ep",
       "config_name": "demon_attack_200ep__demon_attack_fix_latency_0_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -178,7 +183,7 @@
       "source_prefix": "training/demon_attack_200ep/demon_attack_fix_latency_2_200ep",
       "prefix": "latency-aware/demon-attack/demon_attack_200ep/demon_attack_fix_latency_2_200ep",
       "config_name": "demon_attack_200ep__demon_attack_fix_latency_2_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -189,7 +194,7 @@
       "source_prefix": "training/demon_attack_200ep/demon_attack_fix_latency_4_200ep",
       "prefix": "latency-aware/demon-attack/demon_attack_200ep/demon_attack_fix_latency_4_200ep",
       "config_name": "demon_attack_200ep__demon_attack_fix_latency_4_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -200,7 +205,7 @@
       "source_prefix": "training/demon_attack_200ep/demon_attack_fix_latency_6_200ep",
       "prefix": "latency-aware/demon-attack/demon_attack_200ep/demon_attack_fix_latency_6_200ep",
       "config_name": "demon_attack_200ep__demon_attack_fix_latency_6_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -211,7 +216,7 @@
       "source_prefix": "training/demon_attack_200ep/demon_attack_fix_latency_8_200ep",
       "prefix": "latency-aware/demon-attack/demon_attack_200ep/demon_attack_fix_latency_8_200ep",
       "config_name": "demon_attack_200ep__demon_attack_fix_latency_8_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -222,7 +227,7 @@
       "source_prefix": "training/demon_attack_fix_latency_0/data",
       "prefix": "zero-latency/demon-attack/demon_attack_fix_latency_0/data",
       "config_name": "demon_attack_fix_latency_0",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -233,7 +238,7 @@
       "source_prefix": "training/demon_attack_fix_latency_1/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_1/data",
       "config_name": "demon_attack_fix_latency_1",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -244,7 +249,7 @@
       "source_prefix": "training/demon_attack_fix_latency_10/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_10/data",
       "config_name": "demon_attack_fix_latency_10",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -255,7 +260,7 @@
       "source_prefix": "training/demon_attack_fix_latency_11/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_11/data",
       "config_name": "demon_attack_fix_latency_11",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -266,7 +271,7 @@
       "source_prefix": "training/demon_attack_fix_latency_12/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_12/data",
       "config_name": "demon_attack_fix_latency_12",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -277,7 +282,7 @@
       "source_prefix": "training/demon_attack_fix_latency_13/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_13/data",
       "config_name": "demon_attack_fix_latency_13",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -288,7 +293,7 @@
       "source_prefix": "training/demon_attack_fix_latency_14/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_14/data",
       "config_name": "demon_attack_fix_latency_14",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -299,7 +304,7 @@
       "source_prefix": "training/demon_attack_fix_latency_15/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_15/data",
       "config_name": "demon_attack_fix_latency_15",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -310,7 +315,7 @@
       "source_prefix": "training/demon_attack_fix_latency_2/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_2/data",
       "config_name": "demon_attack_fix_latency_2",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -321,7 +326,7 @@
       "source_prefix": "training/demon_attack_fix_latency_3/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_3/data",
       "config_name": "demon_attack_fix_latency_3",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -332,7 +337,7 @@
       "source_prefix": "training/demon_attack_fix_latency_4/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_4/data",
       "config_name": "demon_attack_fix_latency_4",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -343,7 +348,7 @@
       "source_prefix": "training/demon_attack_fix_latency_5/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_5/data",
       "config_name": "demon_attack_fix_latency_5",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -354,7 +359,7 @@
       "source_prefix": "training/demon_attack_fix_latency_6/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_6/data",
       "config_name": "demon_attack_fix_latency_6",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -365,7 +370,7 @@
       "source_prefix": "training/demon_attack_fix_latency_7/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_7/data",
       "config_name": "demon_attack_fix_latency_7",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -376,7 +381,7 @@
       "source_prefix": "training/demon_attack_fix_latency_8/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_8/data",
       "config_name": "demon_attack_fix_latency_8",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -387,7 +392,7 @@
       "source_prefix": "training/demon_attack_fix_latency_9/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_9/data",
       "config_name": "demon_attack_fix_latency_9",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     }
   ]

experiments/latency-transfer/archive/flappy/README.md

@@ -2,16 +2,16 @@
 
 | Training run | Input and latency convention | Retained checkpoint | State |
 |---|---|---|---|
-| [openvla_flappy_fix_latency_0](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0) | , target latency unit not recorded in source config | steps_5000_model.safetensors, steps_5000_pytorch_model.pt | complete_bundle |
-| [openvla_flappy_fix_latency_0_small](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small) | , target latency unit not recorded in source config | best_model.safetensors, steps_1500_model.safetensors | complete_bundle |
-| [openvla_flappy_fix_latency_1_small](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small) | , target latency unit not recorded in source config | best_model.safetensors, steps_1500_model.safetensors | complete_bundle |
-| [openvla_flappy_fix_latency_1_small_tmp](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp) | , target latency unit not recorded in source config | best_model.safetensors, steps_1500_model.safetensors | complete_bundle |
-| [openvla_flappy_fix_latency_2](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2) | , target latency unit not recorded in source config | steps_5000_model.safetensors, steps_5000_pytorch_model.pt | complete_bundle |
-| [openvla_flappy_fix_latency_2_small](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2_small) | , target latency unit not recorded in source config | best_model.safetensors, steps_1500_model.safetensors | complete_bundle |
-| [openvla_flappy_fix_latency_3_small](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_3_small) | , target latency unit not recorded in source config | best_model.safetensors, steps_1500_model.safetensors | complete_bundle |
-| [openvla_flappy_fix_latency_4_exp](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_4_exp) | , target latency unit not recorded in source config | steps_2500_model.safetensors, steps_2500_pytorch_model.pt | complete_bundle |
-| [openvla_flappy_fix_latency_4_small](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_4_small) | , target latency unit not recorded in source config | best_model.safetensors, steps_1500_model.safetensors | complete_bundle |
-| [openvla_flappy_fix_latency_6](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_6) | , target latency unit not recorded in source config | steps_5000_model.safetensors, steps_5000_pytorch_model.pt | complete_bundle |
+| [openvla_flappy_fix_latency_0](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0) | , target latency unit not recorded in source config | steps_5000_model.safetensors, steps_5000_pytorch_model.pt | complete_bundle |
+| [openvla_flappy_fix_latency_0_small](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small) | , target latency unit not recorded in source config | best_model.safetensors, steps_1500_model.safetensors | complete_bundle |
+| [openvla_flappy_fix_latency_1_small](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small) | , target latency unit not recorded in source config | best_model.safetensors, steps_1500_model.safetensors | complete_bundle |
+| [openvla_flappy_fix_latency_1_small_tmp](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp) | , target latency unit not recorded in source config | best_model.safetensors, steps_1500_model.safetensors | complete_bundle |
+| [openvla_flappy_fix_latency_2](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2) | , target latency unit not recorded in source config | steps_5000_model.safetensors, steps_5000_pytorch_model.pt | complete_bundle |
+| [openvla_flappy_fix_latency_2_small](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2_small) | , target latency unit not recorded in source config | best_model.safetensors, steps_1500_model.safetensors | complete_bundle |
+| [openvla_flappy_fix_latency_3_small](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_3_small) | , target latency unit not recorded in source config | best_model.safetensors, steps_1500_model.safetensors | complete_bundle |
+| [openvla_flappy_fix_latency_4_exp](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_4_exp) | , target latency unit not recorded in source config | steps_2500_model.safetensors, steps_2500_pytorch_model.pt | complete_bundle |
+| [openvla_flappy_fix_latency_4_small](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_4_small) | , target latency unit not recorded in source config | best_model.safetensors, steps_1500_model.safetensors | complete_bundle |
+| [openvla_flappy_fix_latency_6](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_6) | , target latency unit not recorded in source config | steps_5000_model.safetensors, steps_5000_pytorch_model.pt | complete_bundle |
 
 ## Asset references
 

experiments/latency-transfer/archive/flappy/manifest.json

@@ -12,18 +12,23 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/checkpoints/steps_5000_model.safetensors"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/checkpoints/steps_5000_pytorch_model.pt"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_flappy_fix_latency_0",
@@ -89,18 +94,23 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/checkpoints/best_model.safetensors"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/checkpoints/steps_1500_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_flappy_fix_latency_0_small",
@@ -116,7 +126,7 @@
           "training_data_mix": "flappy_train__bridge",
           "dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           },
@@ -160,7 +170,7 @@
           "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
           "training_recipe_dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           }
@@ -177,18 +187,23 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/checkpoints/best_model.safetensors"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/checkpoints/steps_1500_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_flappy_fix_latency_1_small",
@@ -204,7 +219,7 @@
           "training_data_mix": "flappy_train__bridge",
           "dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/flappy/flappy_fix_latency_1_small",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           },
@@ -248,7 +263,7 @@
           "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
           "training_recipe_dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/flappy/flappy_fix_latency_1_small",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           }
@@ -265,18 +280,23 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/checkpoints/best_model.safetensors"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/checkpoints/steps_1500_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_flappy_fix_latency_1_small_tmp",
@@ -343,18 +363,23 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/checkpoints/steps_5000_model.safetensors"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/checkpoints/steps_5000_pytorch_model.pt"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_flappy_fix_latency_2",
@@ -420,18 +445,23 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2_small/checkpoints/best_model.safetensors"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2_small/checkpoints/steps_1500_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2_small"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_flappy_fix_latency_2_small",
@@ -447,7 +477,7 @@
           "training_data_mix": "flappy_train__bridge",
           "dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/flappy/flappy_fix_latency_2_small",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           },
@@ -491,7 +521,7 @@
           "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
           "training_recipe_dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/flappy/flappy_fix_latency_2_small",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           }
@@ -508,18 +538,23 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_3_small/checkpoints/best_model.safetensors"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_3_small/checkpoints/steps_1500_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_3_small"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_flappy_fix_latency_3_small",
@@ -535,7 +570,7 @@
           "training_data_mix": "flappy_train__bridge",
           "dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/flappy/flappy_fix_latency_3_small",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           },
@@ -579,7 +614,7 @@
           "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
           "training_recipe_dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/flappy/flappy_fix_latency_3_small",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           }
@@ -596,18 +631,23 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_4_exp/checkpoints/steps_2500_model.safetensors"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_4_exp/checkpoints/steps_2500_pytorch_model.pt"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_4_exp"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_flappy_fix_latency_4_exp",
@@ -673,18 +713,23 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_4_small/checkpoints/best_model.safetensors"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_4_small/checkpoints/steps_1500_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_4_small"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_flappy_fix_latency_4_small",
@@ -700,7 +745,7 @@
           "training_data_mix": "flappy_train__bridge",
           "dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/flappy/flappy_fix_latency_4_small",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           },
@@ -744,7 +789,7 @@
           "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
           "training_recipe_dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/flappy/flappy_fix_latency_4_small",
             "evidence": "Training config uses the exact source dataset name as its local data-root directory. Original metadata and statistics are retained separately."
           }
@@ -761,18 +806,23 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_6/checkpoints/steps_5000_model.safetensors"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_6/checkpoints/steps_5000_pytorch_model.pt"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_6"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_flappy_fix_latency_6",
@@ -839,7 +889,7 @@
       "source_prefix": "training/flappy_fix_latency_0_200ep_context5",
       "prefix": "zero-latency/flappy/flappy_fix_latency_0_200ep_context5",
       "config_name": "flappy_fix_latency_0_200ep_context5",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -850,7 +900,7 @@
       "source_prefix": "training/flappy_fix_latency_1_200ep_context5",
       "prefix": "latency-aware/flappy/flappy_fix_latency_1_200ep_context5",
       "config_name": "flappy_fix_latency_1_200ep_context5",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -861,7 +911,7 @@
       "source_prefix": "training/flappy_fix_latency_2_200ep_context5",
       "prefix": "latency-aware/flappy/flappy_fix_latency_2_200ep_context5",
       "config_name": "flappy_fix_latency_2_200ep_context5",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -872,7 +922,7 @@
       "source_prefix": "training/flappy_fix_latency_3_200ep_context5",
       "prefix": "latency-aware/flappy/flappy_fix_latency_3_200ep_context5",
       "config_name": "flappy_fix_latency_3_200ep_context5",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -883,7 +933,7 @@
       "source_prefix": "training/flappy_fix_latency_4_200ep_context5",
       "prefix": "latency-aware/flappy/flappy_fix_latency_4_200ep_context5",
       "config_name": "flappy_fix_latency_4_200ep_context5",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -894,7 +944,7 @@
       "source_prefix": "training/flappy_200ep/flappy_fix_latency_0_200ep",
       "prefix": "zero-latency/flappy/flappy_200ep/flappy_fix_latency_0_200ep",
       "config_name": "flappy_200ep__flappy_fix_latency_0_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -905,7 +955,7 @@
       "source_prefix": "training/flappy_200ep/flappy_fix_latency_1_200ep",
       "prefix": "latency-aware/flappy/flappy_200ep/flappy_fix_latency_1_200ep",
       "config_name": "flappy_200ep__flappy_fix_latency_1_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -916,7 +966,7 @@
       "source_prefix": "training/flappy_200ep/flappy_fix_latency_2_200ep",
       "prefix": "latency-aware/flappy/flappy_200ep/flappy_fix_latency_2_200ep",
       "config_name": "flappy_200ep__flappy_fix_latency_2_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -927,7 +977,7 @@
       "source_prefix": "training/flappy_200ep/flappy_fix_latency_3_200ep",
       "prefix": "latency-aware/flappy/flappy_200ep/flappy_fix_latency_3_200ep",
       "config_name": "flappy_200ep__flappy_fix_latency_3_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -938,7 +988,7 @@
       "source_prefix": "training/flappy_200ep/flappy_fix_latency_4_200ep",
       "prefix": "latency-aware/flappy/flappy_200ep/flappy_fix_latency_4_200ep",
       "config_name": "flappy_200ep__flappy_fix_latency_4_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -949,7 +999,7 @@
       "source_prefix": "training/flappy_fix_latency_0_small/data",
       "prefix": "zero-latency/flappy/flappy_fix_latency_0_small/data",
       "config_name": "flappy_fix_latency_0_small",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -960,7 +1010,7 @@
       "source_prefix": "training/flappy_fix_latency_1_small/data",
       "prefix": "latency-aware/flappy/flappy_fix_latency_1_small/data",
       "config_name": "flappy_fix_latency_1_small",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -971,7 +1021,7 @@
       "source_prefix": "training/flappy_fix_latency_2_small/data",
       "prefix": "latency-aware/flappy/flappy_fix_latency_2_small/data",
       "config_name": "flappy_fix_latency_2_small",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -982,7 +1032,7 @@
       "source_prefix": "training/flappy_fix_latency_3_small/data",
       "prefix": "latency-aware/flappy/flappy_fix_latency_3_small/data",
       "config_name": "flappy_fix_latency_3_small",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -993,7 +1043,7 @@
       "source_prefix": "training/flappy_fix_latency_4_small/data",
       "prefix": "latency-aware/flappy/flappy_fix_latency_4_small/data",
       "config_name": "flappy_fix_latency_4_small",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     }
   ]

experiments/latency-transfer/demon-attack/README.md

@@ -2,11 +2,11 @@
 
 | Training run | Input and latency convention | Retained checkpoint | State |
 |---|---|---|---|
-| [demon_attack_fix_latency_0_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep) | , target latency unit not recorded in source config | steps_7000_model.safetensors | complete_bundle |
-| [demon_attack_fix_latency_2_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep) | , target latency unit not recorded in source config | steps_7000_model.safetensors | complete_bundle |
-| [demon_attack_fix_latency_4_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep) | , target latency unit not recorded in source config | steps_7000_model.safetensors | complete_bundle |
-| [demon_attack_fix_latency_6_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep) | , target latency unit not recorded in source config | steps_7000_model.safetensors | complete_bundle |
-| [demon_attack_fix_latency_8_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep) | , target latency unit not recorded in source config | steps_7000_model.safetensors | complete_bundle |
+| [demon_attack_fix_latency_0_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep) | , target latency unit not recorded in source config | steps_7000_model.safetensors | complete_bundle |
+| [demon_attack_fix_latency_2_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep) | , target latency unit not recorded in source config | steps_7000_model.safetensors | complete_bundle |
+| [demon_attack_fix_latency_4_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep) | , target latency unit not recorded in source config | steps_7000_model.safetensors | complete_bundle |
+| [demon_attack_fix_latency_6_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep) | , target latency unit not recorded in source config | steps_7000_model.safetensors | complete_bundle |
+| [demon_attack_fix_latency_8_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep) | , target latency unit not recorded in source config | steps_7000_model.safetensors | complete_bundle |
 
 ## Training and evaluation settings
 

experiments/latency-transfer/demon-attack/manifest.json

@@ -12,13 +12,18 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/checkpoints/steps_7000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_demon_attack_200ep",
@@ -70,10 +75,24 @@
             "paper/figures/fig6_latency_robustness/demon_attack"
           ],
           "paper_identity_confirmed": true,
-          "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
+          "interpretation": "The run and checkpoint path are bound to the paper episode CSV and source manifest. Original episode JSONLs/action logs were not recovered. Training-data source limits are recorded separately.",
           "unpublished_training_data": "The recorded local training directory has no recovered exact HF source. No replacement dataset is assumed."
         }
       ],
+      "evaluation_evidence": {
+        "episodes": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/demon-attack/latency-transfer-qwenoft/episodes.csv"
+        },
+        "source_manifest": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/demon-attack/latency-transfer-qwenoft/source_manifest.csv"
+        },
+        "checkpoint_run": "demon_attack_fix_latency_0_200ep",
+        "scope": "Paper run/checkpoint paths and episode-level metrics are retained. Original episode JSONLs and action logs are not recovered."
+      },
       "training_data": []
     },
     {
@@ -85,13 +104,18 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/checkpoints/steps_7000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_demon_attack_200ep",
@@ -143,10 +167,24 @@
             "paper/figures/fig6_latency_robustness/demon_attack"
           ],
           "paper_identity_confirmed": true,
-          "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
+          "interpretation": "The run and checkpoint path are bound to the paper episode CSV and source manifest. Original episode JSONLs/action logs were not recovered. Training-data source limits are recorded separately.",
           "unpublished_training_data": "The recorded local training directory has no recovered exact HF source. No replacement dataset is assumed."
         }
       ],
+      "evaluation_evidence": {
+        "episodes": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/demon-attack/latency-transfer-qwenoft/episodes.csv"
+        },
+        "source_manifest": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/demon-attack/latency-transfer-qwenoft/source_manifest.csv"
+        },
+        "checkpoint_run": "demon_attack_fix_latency_2_200ep",
+        "scope": "Paper run/checkpoint paths and episode-level metrics are retained. Original episode JSONLs and action logs are not recovered."
+      },
       "training_data": []
     },
     {
@@ -158,13 +196,18 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/checkpoints/steps_7000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_demon_attack_200ep",
@@ -216,10 +259,24 @@
             "paper/figures/fig6_latency_robustness/demon_attack"
           ],
           "paper_identity_confirmed": true,
-          "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
+          "interpretation": "The run and checkpoint path are bound to the paper episode CSV and source manifest. Original episode JSONLs/action logs were not recovered. Training-data source limits are recorded separately.",
           "unpublished_training_data": "The recorded local training directory has no recovered exact HF source. No replacement dataset is assumed."
         }
       ],
+      "evaluation_evidence": {
+        "episodes": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/demon-attack/latency-transfer-qwenoft/episodes.csv"
+        },
+        "source_manifest": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/demon-attack/latency-transfer-qwenoft/source_manifest.csv"
+        },
+        "checkpoint_run": "demon_attack_fix_latency_4_200ep",
+        "scope": "Paper run/checkpoint paths and episode-level metrics are retained. Original episode JSONLs and action logs are not recovered."
+      },
       "training_data": []
     },
     {
@@ -231,13 +288,18 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/checkpoints/steps_7000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_demon_attack_200ep",
@@ -289,10 +351,24 @@
             "paper/figures/fig6_latency_robustness/demon_attack"
           ],
           "paper_identity_confirmed": true,
-          "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
+          "interpretation": "The run and checkpoint path are bound to the paper episode CSV and source manifest. Original episode JSONLs/action logs were not recovered. Training-data source limits are recorded separately.",
           "unpublished_training_data": "The recorded local training directory has no recovered exact HF source. No replacement dataset is assumed."
         }
       ],
+      "evaluation_evidence": {
+        "episodes": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/demon-attack/latency-transfer-qwenoft/episodes.csv"
+        },
+        "source_manifest": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/demon-attack/latency-transfer-qwenoft/source_manifest.csv"
+        },
+        "checkpoint_run": "demon_attack_fix_latency_6_200ep",
+        "scope": "Paper run/checkpoint paths and episode-level metrics are retained. Original episode JSONLs and action logs are not recovered."
+      },
       "training_data": []
     },
     {
@@ -304,13 +380,18 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/checkpoints/steps_7000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_demon_attack_200ep",
@@ -362,10 +443,24 @@
             "paper/figures/fig6_latency_robustness/demon_attack"
           ],
           "paper_identity_confirmed": true,
-          "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
+          "interpretation": "The run and checkpoint path are bound to the paper episode CSV and source manifest. Original episode JSONLs/action logs were not recovered. Training-data source limits are recorded separately.",
           "unpublished_training_data": "The recorded local training directory has no recovered exact HF source. No replacement dataset is assumed."
         }
       ],
+      "evaluation_evidence": {
+        "episodes": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/demon-attack/latency-transfer-qwenoft/episodes.csv"
+        },
+        "source_manifest": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/demon-attack/latency-transfer-qwenoft/source_manifest.csv"
+        },
+        "checkpoint_run": "demon_attack_fix_latency_8_200ep",
+        "scope": "Paper run/checkpoint paths and episode-level metrics are retained. Original episode JSONLs and action logs are not recovered."
+      },
       "training_data": []
     }
   ],
@@ -378,7 +473,7 @@
       "source_prefix": "training/demon_attack_200ep/demon_attack_fix_latency_0_200ep",
       "prefix": "zero-latency/demon-attack/demon_attack_200ep/demon_attack_fix_latency_0_200ep",
       "config_name": "demon_attack_200ep__demon_attack_fix_latency_0_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -389,7 +484,7 @@
       "source_prefix": "training/demon_attack_200ep/demon_attack_fix_latency_2_200ep",
       "prefix": "latency-aware/demon-attack/demon_attack_200ep/demon_attack_fix_latency_2_200ep",
       "config_name": "demon_attack_200ep__demon_attack_fix_latency_2_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -400,7 +495,7 @@
       "source_prefix": "training/demon_attack_200ep/demon_attack_fix_latency_4_200ep",
       "prefix": "latency-aware/demon-attack/demon_attack_200ep/demon_attack_fix_latency_4_200ep",
       "config_name": "demon_attack_200ep__demon_attack_fix_latency_4_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -411,7 +506,7 @@
       "source_prefix": "training/demon_attack_200ep/demon_attack_fix_latency_6_200ep",
       "prefix": "latency-aware/demon-attack/demon_attack_200ep/demon_attack_fix_latency_6_200ep",
       "config_name": "demon_attack_200ep__demon_attack_fix_latency_6_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -422,7 +517,7 @@
       "source_prefix": "training/demon_attack_200ep/demon_attack_fix_latency_8_200ep",
       "prefix": "latency-aware/demon-attack/demon_attack_200ep/demon_attack_fix_latency_8_200ep",
       "config_name": "demon_attack_200ep__demon_attack_fix_latency_8_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -433,7 +528,7 @@
       "source_prefix": "training/demon_attack_fix_latency_0/data",
       "prefix": "zero-latency/demon-attack/demon_attack_fix_latency_0/data",
       "config_name": "demon_attack_fix_latency_0",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -444,7 +539,7 @@
       "source_prefix": "training/demon_attack_fix_latency_1/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_1/data",
       "config_name": "demon_attack_fix_latency_1",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -455,7 +550,7 @@
       "source_prefix": "training/demon_attack_fix_latency_10/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_10/data",
       "config_name": "demon_attack_fix_latency_10",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -466,7 +561,7 @@
       "source_prefix": "training/demon_attack_fix_latency_11/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_11/data",
       "config_name": "demon_attack_fix_latency_11",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -477,7 +572,7 @@
       "source_prefix": "training/demon_attack_fix_latency_12/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_12/data",
       "config_name": "demon_attack_fix_latency_12",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -488,7 +583,7 @@
       "source_prefix": "training/demon_attack_fix_latency_13/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_13/data",
       "config_name": "demon_attack_fix_latency_13",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -499,7 +594,7 @@
       "source_prefix": "training/demon_attack_fix_latency_14/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_14/data",
       "config_name": "demon_attack_fix_latency_14",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -510,7 +605,7 @@
       "source_prefix": "training/demon_attack_fix_latency_15/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_15/data",
       "config_name": "demon_attack_fix_latency_15",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -521,7 +616,7 @@
       "source_prefix": "training/demon_attack_fix_latency_2/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_2/data",
       "config_name": "demon_attack_fix_latency_2",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -532,7 +627,7 @@
       "source_prefix": "training/demon_attack_fix_latency_3/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_3/data",
       "config_name": "demon_attack_fix_latency_3",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -543,7 +638,7 @@
       "source_prefix": "training/demon_attack_fix_latency_4/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_4/data",
       "config_name": "demon_attack_fix_latency_4",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -554,7 +649,7 @@
       "source_prefix": "training/demon_attack_fix_latency_5/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_5/data",
       "config_name": "demon_attack_fix_latency_5",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -565,7 +660,7 @@
       "source_prefix": "training/demon_attack_fix_latency_6/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_6/data",
       "config_name": "demon_attack_fix_latency_6",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -576,7 +671,7 @@
       "source_prefix": "training/demon_attack_fix_latency_7/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_7/data",
       "config_name": "demon_attack_fix_latency_7",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -587,7 +682,7 @@
       "source_prefix": "training/demon_attack_fix_latency_8/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_8/data",
       "config_name": "demon_attack_fix_latency_8",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -598,7 +693,7 @@
       "source_prefix": "training/demon_attack_fix_latency_9/data",
       "prefix": "latency-aware/demon-attack/demon_attack_fix_latency_9/data",
       "config_name": "demon_attack_fix_latency_9",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     }
   ]

experiments/latency-transfer/flappy/README.md

@@ -2,18 +2,18 @@
 
 | Training run | Input and latency convention | Retained checkpoint | State |
 |---|---|---|---|
-| [flappy_curriculum_cumulative_40ep_per_latency](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency) | , target latency unit not recorded in source config | steps_5000_model.safetensors | complete_bundle |
-| [flappy_curriculum_exclusive_40ep_per_latency](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency) | , target latency unit not recorded in source config | steps_5000_model.safetensors | complete_bundle |
-| [flappy_fix_latency_0_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep) | , target latency unit not recorded in source config | steps_5000_model.safetensors | complete_bundle |
-| [flappy_fix_latency_1_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep) | , target latency unit not recorded in source config | steps_5000_model.safetensors | complete_bundle |
-| [flappy_fix_latency_2_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep) | , target latency unit not recorded in source config | steps_5000_model.safetensors | complete_bundle |
-| [flappy_fix_latency_3_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep) | , target latency unit not recorded in source config | steps_5000_model.safetensors | complete_bundle |
-| [flappy_fix_latency_4_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep) | , target latency unit not recorded in source config | steps_5000_model.safetensors | complete_bundle |
-| [wan_oft_flappy_fix_latency_0_context5_2000_effbs128](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/zero-latency/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_0_context5_2000_effbs128) | image_mode=single, num_obs_frames=1, target latency unit not recorded in source config | steps_2000_pytorch_model.pt | complete_bundle |
-| [wan_oft_flappy_fix_latency_1_context5_2000_effbs128](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_1_context5_2000_effbs128) | image_mode=single, num_obs_frames=1, target latency unit not recorded in source config | steps_2000_pytorch_model.pt | complete_bundle |
-| [wan_oft_flappy_fix_latency_2_context5_2000_effbs128](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_2_context5_2000_effbs128) | image_mode=single, num_obs_frames=1, target latency unit not recorded in source config | steps_2000_pytorch_model.pt | complete_bundle |
-| [wan_oft_flappy_fix_latency_3_context5_2000_effbs128](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_2000_effbs128) | image_mode=single, num_obs_frames=1, target latency unit not recorded in source config | steps_2000_pytorch_model.pt | complete_bundle |
-| [wan_oft_flappy_fix_latency_4_context5_2000_effbs128](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_4_context5_2000_effbs128) | image_mode=single, num_obs_frames=1, target latency unit not recorded in source config | steps_2000_pytorch_model.pt | complete_bundle |
+| [flappy_curriculum_cumulative_40ep_per_latency](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency) | , target latency unit not recorded in source config | steps_5000_model.safetensors | complete_bundle |
+| [flappy_curriculum_exclusive_40ep_per_latency](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency) | , target latency unit not recorded in source config | steps_5000_model.safetensors | complete_bundle |
+| [flappy_fix_latency_0_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep) | , target latency unit not recorded in source config | steps_5000_model.safetensors | complete_bundle |
+| [flappy_fix_latency_1_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep) | , target latency unit not recorded in source config | steps_5000_model.safetensors | complete_bundle |
+| [flappy_fix_latency_2_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep) | , target latency unit not recorded in source config | steps_5000_model.safetensors | complete_bundle |
+| [flappy_fix_latency_3_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep) | , target latency unit not recorded in source config | steps_5000_model.safetensors | complete_bundle |
+| [flappy_fix_latency_4_200ep](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep) | , target latency unit not recorded in source config | steps_5000_model.safetensors | complete_bundle |
+| [wan_oft_flappy_fix_latency_0_context5_2000_effbs128](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/zero-latency/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_0_context5_2000_effbs128) | image_mode=single, num_obs_frames=1, target latency unit not recorded in source config | steps_2000_pytorch_model.pt | complete_bundle |
+| [wan_oft_flappy_fix_latency_1_context5_2000_effbs128](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_1_context5_2000_effbs128) | image_mode=single, num_obs_frames=1, target latency unit not recorded in source config | steps_2000_pytorch_model.pt | complete_bundle |
+| [wan_oft_flappy_fix_latency_2_context5_2000_effbs128](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_2_context5_2000_effbs128) | image_mode=single, num_obs_frames=1, target latency unit not recorded in source config | steps_2000_pytorch_model.pt | complete_bundle |
+| [wan_oft_flappy_fix_latency_3_context5_2000_effbs128](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_2000_effbs128) | image_mode=single, num_obs_frames=1, target latency unit not recorded in source config | steps_2000_pytorch_model.pt | complete_bundle |
+| [wan_oft_flappy_fix_latency_4_context5_2000_effbs128](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_4_context5_2000_effbs128) | image_mode=single, num_obs_frames=1, target latency unit not recorded in source config | steps_2000_pytorch_model.pt | complete_bundle |
 
 ## Training and evaluation settings
 

experiments/latency-transfer/flappy/manifest.json

@@ -12,13 +12,18 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_5000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_flappy_200ep",
@@ -70,10 +75,24 @@
             "paper/figures/fig6_latency_robustness/flappy"
           ],
           "paper_identity_confirmed": true,
-          "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
+          "interpretation": "The run and checkpoint path are bound to the paper episode CSV and source manifest. Original episode JSONLs/action logs were not recovered. Training-data source limits are recorded separately.",
           "unpublished_training_data": "The recorded local training directory has no recovered exact HF source. No replacement dataset is assumed."
         }
       ],
+      "evaluation_evidence": {
+        "episodes": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/flappy/latency-transfer-qwenoft/episodes.csv"
+        },
+        "source_manifest": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/flappy/latency-transfer-qwenoft/source_manifest.csv"
+        },
+        "checkpoint_run": "flappy_curriculum_cumulative_40ep_per_latency",
+        "scope": "Paper run/checkpoint paths and episode-level metrics are retained. Original episode JSONLs and action logs are not recovered."
+      },
       "training_data": []
     },
     {
@@ -85,13 +104,18 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_5000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_flappy_200ep",
@@ -143,10 +167,24 @@
             "paper/figures/fig6_latency_robustness/flappy"
           ],
           "paper_identity_confirmed": true,
-          "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
+          "interpretation": "The run and checkpoint path are bound to the paper episode CSV and source manifest. Original episode JSONLs/action logs were not recovered. Training-data source limits are recorded separately.",
           "unpublished_training_data": "The recorded local training directory has no recovered exact HF source. No replacement dataset is assumed."
         }
       ],
+      "evaluation_evidence": {
+        "episodes": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/flappy/latency-transfer-qwenoft/episodes.csv"
+        },
+        "source_manifest": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/flappy/latency-transfer-qwenoft/source_manifest.csv"
+        },
+        "checkpoint_run": "flappy_curriculum_exclusive_40ep_per_latency",
+        "scope": "Paper run/checkpoint paths and episode-level metrics are retained. Original episode JSONLs and action logs are not recovered."
+      },
       "training_data": []
     },
     {
@@ -158,13 +196,18 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/checkpoints/steps_5000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_flappy_200ep",
@@ -214,10 +257,24 @@
             "paper/figures/fig6_latency_robustness/flappy"
           ],
           "paper_identity_confirmed": true,
-          "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
+          "interpretation": "The run and checkpoint path are bound to the paper episode CSV and source manifest. Original episode JSONLs/action logs were not recovered. Training-data source limits are recorded separately.",
           "unpublished_training_data": "The recorded local training directory has no recovered exact HF source. No replacement dataset is assumed."
         }
       ],
+      "evaluation_evidence": {
+        "episodes": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/flappy/latency-transfer-qwenoft/episodes.csv"
+        },
+        "source_manifest": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/flappy/latency-transfer-qwenoft/source_manifest.csv"
+        },
+        "checkpoint_run": "flappy_fix_latency_0_200ep",
+        "scope": "Paper run/checkpoint paths and episode-level metrics are retained. Original episode JSONLs and action logs are not recovered."
+      },
       "training_data": []
     },
     {
@@ -229,13 +286,18 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/checkpoints/steps_5000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_flappy_200ep",
@@ -285,10 +347,24 @@
             "paper/figures/fig6_latency_robustness/flappy"
           ],
           "paper_identity_confirmed": true,
-          "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
+          "interpretation": "The run and checkpoint path are bound to the paper episode CSV and source manifest. Original episode JSONLs/action logs were not recovered. Training-data source limits are recorded separately.",
           "unpublished_training_data": "The recorded local training directory has no recovered exact HF source. No replacement dataset is assumed."
         }
       ],
+      "evaluation_evidence": {
+        "episodes": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/flappy/latency-transfer-qwenoft/episodes.csv"
+        },
+        "source_manifest": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/flappy/latency-transfer-qwenoft/source_manifest.csv"
+        },
+        "checkpoint_run": "flappy_fix_latency_1_200ep",
+        "scope": "Paper run/checkpoint paths and episode-level metrics are retained. Original episode JSONLs and action logs are not recovered."
+      },
       "training_data": []
     },
     {
@@ -300,13 +376,18 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/checkpoints/steps_5000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_flappy_200ep",
@@ -357,10 +438,24 @@
             "paper/figures/fig6_latency_robustness/flappy"
           ],
           "paper_identity_confirmed": true,
-          "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
+          "interpretation": "The run and checkpoint path are bound to the paper episode CSV and source manifest. Original episode JSONLs/action logs were not recovered. Training-data source limits are recorded separately.",
           "unpublished_training_data": "The recorded local training directory has no recovered exact HF source. No replacement dataset is assumed."
         }
       ],
+      "evaluation_evidence": {
+        "episodes": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/flappy/latency-transfer-qwenoft/episodes.csv"
+        },
+        "source_manifest": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/flappy/latency-transfer-qwenoft/source_manifest.csv"
+        },
+        "checkpoint_run": "flappy_fix_latency_2_200ep",
+        "scope": "Paper run/checkpoint paths and episode-level metrics are retained. Original episode JSONLs and action logs are not recovered."
+      },
       "training_data": []
     },
     {
@@ -372,13 +467,18 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/checkpoints/steps_5000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_flappy_200ep",
@@ -429,10 +529,24 @@
             "paper/figures/fig6_latency_robustness/flappy"
           ],
           "paper_identity_confirmed": true,
-          "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
+          "interpretation": "The run and checkpoint path are bound to the paper episode CSV and source manifest. Original episode JSONLs/action logs were not recovered. Training-data source limits are recorded separately.",
           "unpublished_training_data": "The recorded local training directory has no recovered exact HF source. No replacement dataset is assumed."
         }
       ],
+      "evaluation_evidence": {
+        "episodes": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/flappy/latency-transfer-qwenoft/episodes.csv"
+        },
+        "source_manifest": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/flappy/latency-transfer-qwenoft/source_manifest.csv"
+        },
+        "checkpoint_run": "flappy_fix_latency_3_200ep",
+        "scope": "Paper run/checkpoint paths and episode-level metrics are retained. Original episode JSONLs and action logs are not recovered."
+      },
       "training_data": []
     },
     {
@@ -444,13 +558,18 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/checkpoints/steps_5000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep"
+      },
       "evaluation": [],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_flappy_200ep",
@@ -501,10 +620,24 @@
             "paper/figures/fig6_latency_robustness/flappy"
           ],
           "paper_identity_confirmed": true,
-          "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
+          "interpretation": "The run and checkpoint path are bound to the paper episode CSV and source manifest. Original episode JSONLs/action logs were not recovered. Training-data source limits are recorded separately.",
           "unpublished_training_data": "The recorded local training directory has no recovered exact HF source. No replacement dataset is assumed."
         }
       ],
+      "evaluation_evidence": {
+        "episodes": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/flappy/latency-transfer-qwenoft/episodes.csv"
+        },
+        "source_manifest": {
+          "repo": "latency-sensitive-bench/benchmark-datasets",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
+          "path": "evaluations/flappy/latency-transfer-qwenoft/source_manifest.csv"
+        },
+        "checkpoint_run": "flappy_fix_latency_4_200ep",
+        "scope": "Paper run/checkpoint paths and episode-level metrics are retained. Original episode JSONLs and action logs are not recovered."
+      },
       "training_data": []
     },
     {
@@ -516,19 +649,24 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "zero-latency/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_0_context5_2000_effbs128/checkpoints/steps_2000_pytorch_model.pt"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "zero-latency/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_0_context5_2000_effbs128"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/flappy/wan_oft_flappy_fix_latency_0_context5_2000_effbs128/eval/post_train/step_2000.json"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "run": "wan_oft_flappy_fix_latency_0_context5_2000_effbs128",
@@ -544,7 +682,7 @@
           "source_model_revision": "1ba5d2f489f03f9cb07a637ca680fa4f2f8e51a4",
           "dataset": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "zero-latency/flappy/flappy_fix_latency_0_200ep_context5",
             "source_repo": "latency-sensitive-bench/flappy_200ep_context5",
             "source_revision": "19d800e5f7ec9d5311f508dbfe4a50463261616c",
@@ -562,7 +700,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "zero-latency/flappy/flappy_fix_latency_0_200ep_context5"
         }
       ]
@@ -576,39 +714,44 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_1_context5_2000_effbs128/checkpoints/steps_2000_pytorch_model.pt"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_1_context5_2000_effbs128"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/flappy/wan_oft_flappy_fix_latency_1_context5_2000_effbs128/eval/mid_train/step_1000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/flappy/wan_oft_flappy_fix_latency_1_context5_2000_effbs128/eval/mid_train/step_1500.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/flappy/wan_oft_flappy_fix_latency_1_context5_2000_effbs128/eval/mid_train/step_2000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/flappy/wan_oft_flappy_fix_latency_1_context5_2000_effbs128/eval/mid_train/step_500.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/flappy/wan_oft_flappy_fix_latency_1_context5_2000_effbs128/eval/post_train/step_2000.json"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "run": "wan_oft_flappy_fix_latency_1_context5_2000_effbs128",
@@ -624,7 +767,7 @@
           "source_model_revision": "1ba5d2f489f03f9cb07a637ca680fa4f2f8e51a4",
           "dataset": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/flappy/flappy_fix_latency_1_200ep_context5",
             "source_repo": "latency-sensitive-bench/flappy_200ep_context5",
             "source_revision": "19d800e5f7ec9d5311f508dbfe4a50463261616c",
@@ -642,7 +785,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/flappy/flappy_fix_latency_1_200ep_context5"
         }
       ]
@@ -656,19 +799,24 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_2_context5_2000_effbs128/checkpoints/steps_2000_pytorch_model.pt"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_2_context5_2000_effbs128"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/flappy/wan_oft_flappy_fix_latency_2_context5_2000_effbs128/eval/post_train/step_2000.json"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "run": "wan_oft_flappy_fix_latency_2_context5_2000_effbs128",
@@ -684,7 +832,7 @@
           "source_model_revision": "1ba5d2f489f03f9cb07a637ca680fa4f2f8e51a4",
           "dataset": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/flappy/flappy_fix_latency_2_200ep_context5",
             "source_repo": "latency-sensitive-bench/flappy_200ep_context5",
             "source_revision": "19d800e5f7ec9d5311f508dbfe4a50463261616c",
@@ -704,7 +852,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/flappy/flappy_fix_latency_2_200ep_context5"
         }
       ]
@@ -718,39 +866,44 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_2000_effbs128/checkpoints/steps_2000_pytorch_model.pt"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_2000_effbs128"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/flappy/wan_oft_flappy_fix_latency_3_context5_2000_effbs128/eval/mid_train/step_1000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/flappy/wan_oft_flappy_fix_latency_3_context5_2000_effbs128/eval/mid_train/step_1500.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/flappy/wan_oft_flappy_fix_latency_3_context5_2000_effbs128/eval/mid_train/step_2000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/flappy/wan_oft_flappy_fix_latency_3_context5_2000_effbs128/eval/mid_train/step_500.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/flappy/wan_oft_flappy_fix_latency_3_context5_2000_effbs128/eval/post_train/step_2000.json"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "run": "wan_oft_flappy_fix_latency_3_context5_2000_effbs128",
@@ -766,7 +919,7 @@
           "source_model_revision": "1ba5d2f489f03f9cb07a637ca680fa4f2f8e51a4",
           "dataset": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/flappy/flappy_fix_latency_3_200ep_context5",
             "source_repo": "latency-sensitive-bench/flappy_200ep_context5",
             "source_revision": "19d800e5f7ec9d5311f508dbfe4a50463261616c",
@@ -784,7 +937,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/flappy/flappy_fix_latency_3_200ep_context5"
         }
       ]
@@ -798,39 +951,44 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_4_context5_2000_effbs128/checkpoints/steps_2000_pytorch_model.pt"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_4_context5_2000_effbs128"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/flappy/wan_oft_flappy_fix_latency_4_context5_2000_effbs128/eval/mid_train/step_1000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/flappy/wan_oft_flappy_fix_latency_4_context5_2000_effbs128/eval/mid_train/step_1500.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/flappy/wan_oft_flappy_fix_latency_4_context5_2000_effbs128/eval/mid_train/step_2000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/flappy/wan_oft_flappy_fix_latency_4_context5_2000_effbs128/eval/mid_train/step_500.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/flappy/wan_oft_flappy_fix_latency_4_context5_2000_effbs128/eval/post_train/step_2000.json"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "run": "wan_oft_flappy_fix_latency_4_context5_2000_effbs128",
@@ -846,7 +1004,7 @@
           "source_model_revision": "1ba5d2f489f03f9cb07a637ca680fa4f2f8e51a4",
           "dataset": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/flappy/flappy_fix_latency_4_200ep_context5",
             "source_repo": "latency-sensitive-bench/flappy_200ep_context5",
             "source_revision": "19d800e5f7ec9d5311f508dbfe4a50463261616c",
@@ -864,7 +1022,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/flappy/flappy_fix_latency_4_200ep_context5"
         }
       ]
@@ -879,7 +1037,7 @@
       "source_prefix": "training/flappy_fix_latency_0_200ep_context5",
       "prefix": "zero-latency/flappy/flappy_fix_latency_0_200ep_context5",
       "config_name": "flappy_fix_latency_0_200ep_context5",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -890,7 +1048,7 @@
       "source_prefix": "training/flappy_fix_latency_1_200ep_context5",
       "prefix": "latency-aware/flappy/flappy_fix_latency_1_200ep_context5",
       "config_name": "flappy_fix_latency_1_200ep_context5",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -901,7 +1059,7 @@
       "source_prefix": "training/flappy_fix_latency_2_200ep_context5",
       "prefix": "latency-aware/flappy/flappy_fix_latency_2_200ep_context5",
       "config_name": "flappy_fix_latency_2_200ep_context5",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -912,7 +1070,7 @@
       "source_prefix": "training/flappy_fix_latency_3_200ep_context5",
       "prefix": "latency-aware/flappy/flappy_fix_latency_3_200ep_context5",
       "config_name": "flappy_fix_latency_3_200ep_context5",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -923,7 +1081,7 @@
       "source_prefix": "training/flappy_fix_latency_4_200ep_context5",
       "prefix": "latency-aware/flappy/flappy_fix_latency_4_200ep_context5",
       "config_name": "flappy_fix_latency_4_200ep_context5",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -934,7 +1092,7 @@
       "source_prefix": "training/flappy_200ep/flappy_fix_latency_0_200ep",
       "prefix": "zero-latency/flappy/flappy_200ep/flappy_fix_latency_0_200ep",
       "config_name": "flappy_200ep__flappy_fix_latency_0_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -945,7 +1103,7 @@
       "source_prefix": "training/flappy_200ep/flappy_fix_latency_1_200ep",
       "prefix": "latency-aware/flappy/flappy_200ep/flappy_fix_latency_1_200ep",
       "config_name": "flappy_200ep__flappy_fix_latency_1_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -956,7 +1114,7 @@
       "source_prefix": "training/flappy_200ep/flappy_fix_latency_2_200ep",
       "prefix": "latency-aware/flappy/flappy_200ep/flappy_fix_latency_2_200ep",
       "config_name": "flappy_200ep__flappy_fix_latency_2_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -967,7 +1125,7 @@
       "source_prefix": "training/flappy_200ep/flappy_fix_latency_3_200ep",
       "prefix": "latency-aware/flappy/flappy_200ep/flappy_fix_latency_3_200ep",
       "config_name": "flappy_200ep__flappy_fix_latency_3_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -978,7 +1136,7 @@
       "source_prefix": "training/flappy_200ep/flappy_fix_latency_4_200ep",
       "prefix": "latency-aware/flappy/flappy_200ep/flappy_fix_latency_4_200ep",
       "config_name": "flappy_200ep__flappy_fix_latency_4_200ep",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -989,7 +1147,7 @@
       "source_prefix": "training/flappy_fix_latency_0_small/data",
       "prefix": "zero-latency/flappy/flappy_fix_latency_0_small/data",
       "config_name": "flappy_fix_latency_0_small",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -1000,7 +1158,7 @@
       "source_prefix": "training/flappy_fix_latency_1_small/data",
       "prefix": "latency-aware/flappy/flappy_fix_latency_1_small/data",
       "config_name": "flappy_fix_latency_1_small",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -1011,7 +1169,7 @@
       "source_prefix": "training/flappy_fix_latency_2_small/data",
       "prefix": "latency-aware/flappy/flappy_fix_latency_2_small/data",
       "config_name": "flappy_fix_latency_2_small",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -1022,7 +1180,7 @@
       "source_prefix": "training/flappy_fix_latency_3_small/data",
       "prefix": "latency-aware/flappy/flappy_fix_latency_3_small/data",
       "config_name": "flappy_fix_latency_3_small",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -1033,7 +1191,7 @@
       "source_prefix": "training/flappy_fix_latency_4_small/data",
       "prefix": "latency-aware/flappy/flappy_fix_latency_4_small/data",
       "config_name": "flappy_fix_latency_4_small",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     }
   ]

experiments/mean-vs-profile/ant/README.md

@@ -2,8 +2,8 @@
 
 | Condition | Checkpoint | Training latency | Evaluation episodes | Mean return |
 |---|---|---|---|---|
-| Mean-trained | [step 5000](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/ant/vla/starvla-qwenoft-h1/ant-qwenoft-mean-3090-s42) | 90.56460639 ms fixed | 100 | 1453.844063807972 |
-| Profile-trained | [profile-trained](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/ant/vla/starvla-qwenoft-h1/ant_rgb_state_profile_20260911T112232Z_openvla_native_sft_5k) | Temporal profile | 20 | 1853.3319204046265 |
+| Mean-trained | [step 5000](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/ant/vla/starvla-qwenoft-h1/ant-qwenoft-mean-3090-s42) | 90.56460639 ms fixed | 100 | 1453.844063807972 |
+| Profile-trained | [profile-trained](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/ant/vla/starvla-qwenoft-h1/ant_rgb_state_profile_20260911T112232Z_openvla_native_sft_5k) | Temporal profile | 20 | 1853.3319204046265 |
 
 ## Evaluation protocol
 

experiments/mean-vs-profile/ant/manifest.json

@@ -18,7 +18,7 @@
         "task": "ant",
         "label": "Mean-trained training data",
         "files": 27,
-        "revision": "15d17287246bd19940c837d6c8a5327564f969c5"
+        "revision": "27d2fe8badf83302809f5570e3118bd987f15600"
       },
       "fixed_training_latency_ms": 90.56460638563993,
       "evaluation_profile": {
@@ -31,11 +31,11 @@
         "label": "Evaluation temporal profile",
         "files": 3,
         "sha256": "0e49f72ec05ac600a54e07bbca5af3143708444d606167f33fa21788a9fefe50",
-        "revision": "e7371fae25471e7b2052819289ca3d5093e85559"
+        "revision": "dc603c91680f78ae3256e5559758cbee89864ce0"
       },
       "evaluation": {
         "repo": "latency-sensitive-bench/benchmark-datasets",
-        "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+        "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
         "path": "evaluations/ant/ant-qwenoft-mean-3090-s42/evaluation/profile-simulation-100ep-20261001"
       },
       "result": {
@@ -103,12 +103,12 @@
       "episodes": 20,
       "evaluation": {
         "repo": "latency-sensitive-bench/benchmark-datasets",
-        "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+        "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
         "path": "evaluations/ant/ant_rgb_state_profile_20260911T112232Z_openvla_native_sft_5k/hardware-20ep/summary.json"
       },
       "training_data": {
         "repo": "latency-sensitive-bench/benchmark-datasets",
-        "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+        "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
         "path": "latency-aware/ant/ant_rgb_state_profile_20260911T112232Z/raw"
       }
     }

experiments/mean-vs-profile/deadly-corridor/README.md

@@ -2,8 +2,8 @@
 
 | Condition | Checkpoint | Training latency | Evaluation episodes | Mean return |
 |---|---|---|---|---|
-| Mean-trained | [step 5000](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor-qwenoft-mean-3090-s42) | 73.69250777 ms fixed | 100 | 1620.7987757873534 |
-| Profile-trained | [profile-trained](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor-qwenoft-hist8-scratch-s42-v2) | Temporal profile | 100 | 2091.234108123779 |
+| Mean-trained | [step 5000](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor-qwenoft-mean-3090-s42) | 73.69250777 ms fixed | 100 | 1620.7987757873534 |
+| Profile-trained | [profile-trained](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor-qwenoft-hist8-scratch-s42-v2) | Temporal profile | 100 | 2091.234108123779 |
 
 ## Evaluation protocol
 

experiments/mean-vs-profile/deadly-corridor/manifest.json

@@ -18,7 +18,7 @@
         "task": "deadly-corridor",
         "label": "Mean-trained training data",
         "files": 27,
-        "revision": "15d17287246bd19940c837d6c8a5327564f969c5"
+        "revision": "27d2fe8badf83302809f5570e3118bd987f15600"
       },
       "fixed_training_latency_ms": 73.69250777493353,
       "evaluation_profile": {
@@ -31,11 +31,11 @@
         "label": "Evaluation temporal profile",
         "files": 3,
         "sha256": "bbfb93cef9f63750a9a159ec10a93a9fc6c25e4cb0cac3b39532663b4dc11eba",
-        "revision": "e7371fae25471e7b2052819289ca3d5093e85559"
+        "revision": "dc603c91680f78ae3256e5559758cbee89864ce0"
       },
       "evaluation": {
         "repo": "latency-sensitive-bench/benchmark-datasets",
-        "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+        "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
         "path": "evaluations/deadly-corridor/deadly_corridor-qwenoft-mean-3090-s42/evaluation/profile-simulation-100ep-20261001"
       },
       "result": {
@@ -105,14 +105,14 @@
         "label": "Evaluation temporal profile",
         "files": 3,
         "sha256": "bbfb93cef9f63750a9a159ec10a93a9fc6c25e4cb0cac3b39532663b4dc11eba",
-        "revision": "e7371fae25471e7b2052819289ca3d5093e85559"
+        "revision": "dc603c91680f78ae3256e5559758cbee89864ce0"
       },
       "mean_return": 2091.234108123779,
       "std_return": 579.6123067681043,
       "episodes": 100,
       "evaluation": {
         "repo": "latency-sensitive-bench/benchmark-datasets",
-        "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+        "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
         "path": "evaluations/deadly-corridor/deadly_corridor-qwenoft-hist8-scratch-s42-v2"
       }
     }

experiments/mean-vs-profile/flappy/README.md

@@ -2,8 +2,8 @@
 
 | Condition | Checkpoint | Training latency | Evaluation episodes | Mean return |
 |---|---|---|---|---|
-| Mean-trained | [step 5000](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy-qwenoft-mean-3090-s42) | 75.87417451 ms fixed | 100 | 384.8240045265853 |
-| Profile-trained | [profile-trained](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy-qwenoft-hist8-scratch-s42-v2) | Temporal profile | 100 | 373.7790043977648 |
+| Mean-trained | [step 5000](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy-qwenoft-mean-3090-s42) | 75.87417451 ms fixed | 100 | 384.8240045265853 |
+| Profile-trained | [profile-trained](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy-qwenoft-hist8-scratch-s42-v2) | Temporal profile | 100 | 373.7790043977648 |
 
 ## Evaluation protocol
 

experiments/mean-vs-profile/flappy/manifest.json

@@ -18,7 +18,7 @@
         "task": "flappy",
         "label": "Mean-trained training data",
         "files": 30,
-        "revision": "15d17287246bd19940c837d6c8a5327564f969c5"
+        "revision": "27d2fe8badf83302809f5570e3118bd987f15600"
       },
       "fixed_training_latency_ms": 75.87417450998383,
       "evaluation_profile": {
@@ -31,11 +31,11 @@
         "label": "Evaluation temporal profile",
         "files": 3,
         "sha256": "c266cba7ebac05c7e37bc83f1f16fe4b27dab97942ff43290e6c41514f6e39fc",
-        "revision": "e7371fae25471e7b2052819289ca3d5093e85559"
+        "revision": "dc603c91680f78ae3256e5559758cbee89864ce0"
       },
       "evaluation": {
         "repo": "latency-sensitive-bench/benchmark-datasets",
-        "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+        "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
         "path": "evaluations/flappy/flappy-qwenoft-mean-3090-s42/evaluation/profile-simulation-100ep-20261001"
       },
       "result": {
@@ -105,14 +105,14 @@
         "label": "Evaluation temporal profile",
         "files": 3,
         "sha256": "c266cba7ebac05c7e37bc83f1f16fe4b27dab97942ff43290e6c41514f6e39fc",
-        "revision": "e7371fae25471e7b2052819289ca3d5093e85559"
+        "revision": "dc603c91680f78ae3256e5559758cbee89864ce0"
       },
       "mean_return": 373.7790043977648,
       "std_return": 126.82699726383998,
       "episodes": 100,
       "evaluation": {
         "repo": "latency-sensitive-bench/benchmark-datasets",
-        "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+        "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
         "path": "evaluations/flappy/flappy-qwenoft-hist8-scratch-s42-v2"
       }
     }

experiments/mean-vs-profile/intercept-grab-fast/README.md

@@ -2,8 +2,8 @@
 
 | Condition | Checkpoint | Training latency | Evaluation episodes | Mean return |
 |---|---|---|---|---|
-| Mean-trained | [step 5000](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/mikasa-intercept-grab-fast/vla/starvla-qwenoft-h1/intercept-qwenoft-mean-3090-h1-s0) | 99.05021290 ms fixed | 100 | 3.5443485127069287 |
-| Profile-trained | [profile-trained](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/mikasa-intercept-grab-fast/vla/starvla-qwenoft-h1/mikasa_qwenoft_latency_3090_2xh200_20260922_evalfix_20260923) | Temporal profile | 100 | 18.195560179747673 |
+| Mean-trained | [step 5000](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/mikasa-intercept-grab-fast/vla/starvla-qwenoft-h1/intercept-qwenoft-mean-3090-h1-s0) | 99.05021290 ms fixed | 100 | 3.5443485127069287 |
+| Profile-trained | [profile-trained](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/mikasa-intercept-grab-fast/vla/starvla-qwenoft-h1/mikasa_qwenoft_latency_3090_2xh200_20260922_evalfix_20260923) | Temporal profile | 100 | 18.195560179747673 |
 
 ## Evaluation protocol
 

experiments/mean-vs-profile/intercept-grab-fast/manifest.json

@@ -18,7 +18,7 @@
         "task": "intercept-grab-fast",
         "label": "Mean-trained training data",
         "files": 309,
-        "revision": "15d17287246bd19940c837d6c8a5327564f969c5"
+        "revision": "27d2fe8badf83302809f5570e3118bd987f15600"
       },
       "fixed_training_latency_ms": 99.05021289731565,
       "evaluation_profile": {
@@ -31,11 +31,11 @@
         "label": "Evaluation temporal profile",
         "files": 3,
         "sha256": "31a9b15d6b04182439934f0b377cb3d09be16ce21f3cf95c1049918f8a5d4984",
-        "revision": "e7371fae25471e7b2052819289ca3d5093e85559"
+        "revision": "dc603c91680f78ae3256e5559758cbee89864ce0"
       },
       "evaluation": {
         "repo": "latency-sensitive-bench/benchmark-datasets",
-        "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+        "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
         "path": "evaluations/intercept-grab-fast/intercept-qwenoft-mean-3090-h1-s0/evaluation/profile-simulation-100ep-20261001"
       },
       "result": {
@@ -107,14 +107,14 @@
         "label": "Evaluation temporal profile",
         "files": 3,
         "sha256": "31a9b15d6b04182439934f0b377cb3d09be16ce21f3cf95c1049918f8a5d4984",
-        "revision": "e7371fae25471e7b2052819289ca3d5093e85559"
+        "revision": "dc603c91680f78ae3256e5559758cbee89864ce0"
       },
       "mean_return": 18.195560179747673,
       "std_return": 15.338579798327062,
       "episodes": 100,
       "evaluation": {
         "repo": "latency-sensitive-bench/benchmark-datasets",
-        "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+        "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
         "path": "evaluations/intercept-grab-fast/mikasa_qwenoft_latency_3090_2xh200_20260922_evalfix_20260923"
       }
     }

experiments/visual-history/archive/deadly-corridor/README.md

@@ -2,8 +2,8 @@
 
 | Training run | Input and latency convention | Retained checkpoint | State |
 |---|---|---|---|
-| [deadly_corridor_fix_latency_0_1000ep_7k2steps_single_baseline](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_0_1000ep_7k2steps_single_baseline) | image_mode=single, num_obs_frames=1, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
-| [memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/deadly-corridor/small-policy/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0) | Original teacher configuration | checkpoint_000048836_25004032.pth | teacher_bundle |
+| [deadly_corridor_fix_latency_0_1000ep_7k2steps_single_baseline](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_0_1000ep_7k2steps_single_baseline) | image_mode=single, num_obs_frames=1, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
+| [memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/deadly-corridor/small-policy/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0) | Original teacher configuration | checkpoint_000048836_25004032.pth | teacher_bundle |
 
 ## Asset references
 

experiments/visual-history/archive/deadly-corridor/manifest.json

@@ -12,34 +12,39 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_0_1000ep_7k2steps_single_baseline/checkpoints/steps_4000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_0_1000ep_7k2steps_single_baseline"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_0_1000ep_7k2steps_single_baseline/hydra/.hydra/hydra.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_0_1000ep_7k2steps_single_baseline/hydra/.hydra/overrides.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_0_1000ep_7k2steps_single_baseline/hydra/train_starvla_hydra.log"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_0_1000ep_7k2steps_single_baseline/summary.jsonl"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "run": "deadly_corridor_fix_latency_0_1000ep_7k2steps_single_baseline",
@@ -67,7 +72,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "zero-latency/deadly-corridor/deadly_corridor_fixed_latency_0_1000ep_7k2steps",
           "config_name": "deadly_corridor_fixed_latency_0_1000ep_7k2steps"
         }
@@ -82,29 +87,34 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/deadly-corridor/small-policy/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/train/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/checkpoint_p0/checkpoint_000048836_25004032.pth"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/deadly-corridor/small-policy/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/train/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/episode_metrics.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/train/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/git.diff"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/train/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/sf_log.txt"
         }
       ],
       "asset_state": "teacher_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "run": "memory/rollout_teachers/deadly_corridor",
@@ -132,7 +142,7 @@
       "source_prefix": "deadly_corridor_fixed_latency_6_1000ep_7k2steps",
       "prefix": "latency-aware/deadly-corridor/deadly_corridor_fixed_latency_6_1000ep_7k2steps",
       "config_name": "deadly_corridor_fixed_latency_6_1000ep_7k2steps",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -143,82 +153,8 @@
       "source_prefix": "deadly_corridor_fixed_latency_0_1000ep_7k2steps",
       "prefix": "zero-latency/deadly-corridor/deadly_corridor_fixed_latency_0_1000ep_7k2steps",
       "config_name": "deadly_corridor_fixed_latency_0_1000ep_7k2steps",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     }
-  ],
-  "paper_rows": [
-    {
-      "task_key": "deadly_corridor",
-      "method_key": "single_baseline",
-      "checkpoint_step": "4000",
-      "source_history_step": "4000",
-      "latency_raw_frames": "6",
-      "mean_reward": "232.00475982666015",
-      "std_reward": "257.1936181111822",
-      "mean_length": "64.61",
-      "std_length": "24.343333789766756",
-      "episode_count": "100",
-      "reward_ci95_lo": "180.71487235229466",
-      "reward_ci95_hi": "283.29464730102563",
-      "length_ci95_lo": "59.755420559042356",
-      "length_ci95_hi": "69.46457944095764",
-      "status": "complete",
-      "source_run_id": "sxgh13gg"
-    },
-    {
-      "task_key": "deadly_corridor",
-      "method_key": "plain_multi",
-      "checkpoint_step": "4000",
-      "source_history_step": "4000",
-      "latency_raw_frames": "6",
-      "mean_reward": "227.0985302734375",
-      "std_reward": "170.04333449353234",
-      "mean_length": "66.55",
-      "std_length": "23.632763274742125",
-      "episode_count": "100",
-      "reward_ci95_lo": "193.18826626304144",
-      "reward_ci95_hi": "261.00879428383354",
-      "length_ci95_lo": "61.837123459860294",
-      "length_ci95_hi": "71.2628765401397",
-      "status": "complete",
-      "source_run_id": "g8yprapd"
-    },
-    {
-      "task_key": "deadly_corridor",
-      "method_key": "stitch",
-      "checkpoint_step": "4000",
-      "source_history_step": "4000",
-      "latency_raw_frames": "6",
-      "mean_reward": "271.0697445678711",
-      "std_reward": "284.4073394474786",
-      "mean_length": "79.47",
-      "std_length": "32.63907320988143",
-      "episode_count": "100",
-      "reward_ci95_lo": "214.35286121643844",
-      "reward_ci95_hi": "327.7866279193038",
-      "length_ci95_lo": "72.96107336139764",
-      "length_ci95_hi": "85.97892663860236",
-      "status": "complete",
-      "source_run_id": "6lxzrons"
-    },
-    {
-      "task_key": "deadly_corridor",
-      "method_key": "kv_memory_flex",
-      "checkpoint_step": "4000",
-      "source_history_step": "4000",
-      "latency_raw_frames": "6",
-      "mean_reward": "229.2106620788574",
-      "std_reward": "166.31335369199877",
-      "mean_length": "66.96",
-      "std_length": "22.641077712865172",
-      "episode_count": "100",
-      "reward_ci95_lo": "196.04423571496744",
-      "reward_ci95_hi": "262.3770884427474",
-      "length_ci95_lo": "62.44488669073733",
-      "length_ci95_hi": "71.47511330926267",
-      "status": "complete",
-      "source_run_id": "wb7qhv7z"
-    }
   ]
 }

experiments/visual-history/archive/demon-attack/README.md

@@ -2,12 +2,12 @@
 
 | Training run | Input and latency convention | Retained checkpoint | State |
 |---|---|---|---|
-| [demon_attack_fix_latency_0_200ep_7k2steps_single_baseline](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep_7k2steps_single_baseline) | image_mode=single, num_obs_frames=1, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
-| [openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_model.safetensors | complete_bundle |
-| [openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2) | image_mode=single, num_obs_frames=1, target latency unit=raw_frames | — | metadata_only |
-| [openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_exp2](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_exp2) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_model.safetensors | complete_bundle |
-| [openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_seed1](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_seed1) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_model.safetensors | complete_bundle |
-| [memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/demon-attack/small-policy/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0) | Original teacher configuration | checkpoint_000048880_25026560.pth | teacher_bundle |
+| [demon_attack_fix_latency_0_200ep_7k2steps_single_baseline](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep_7k2steps_single_baseline) | image_mode=single, num_obs_frames=1, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
+| [openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_model.safetensors | complete_bundle |
+| [openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2) | image_mode=single, num_obs_frames=1, target latency unit=raw_frames | — | metadata_only |
+| [openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_exp2](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_exp2) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_model.safetensors | shared_checkpoint |
+| [openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_seed1](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_seed1) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_model.safetensors | shared_checkpoint |
+| [memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/demon-attack/small-policy/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0) | Original teacher configuration | checkpoint_000048880_25026560.pth | teacher_bundle |
 
 ## Asset references
 

experiments/visual-history/archive/demon-attack/manifest.json

@@ -12,34 +12,39 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep_7k2steps_single_baseline/checkpoints/steps_4000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep_7k2steps_single_baseline"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_0_200ep_7k2steps_single_baseline/hydra/.hydra/hydra.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_0_200ep_7k2steps_single_baseline/hydra/.hydra/overrides.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_0_200ep_7k2steps_single_baseline/hydra/train_starvla_hydra.log"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_0_200ep_7k2steps_single_baseline/summary.jsonl"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "run": "demon_attack_fix_latency_0_200ep_7k2steps_single_baseline",
@@ -67,7 +72,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "zero-latency/demon-attack/demon_attack_fixed_latency_0_200ep_7k2steps",
           "config_name": "demon_attack_fixed_latency_0_200ep_7k2steps"
         }
@@ -82,44 +87,49 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15/checkpoints/steps_5000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15/eval/mid_train/step_1000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15/eval/mid_train/step_2000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15/eval/mid_train/step_3000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15/eval/mid_train/step_4000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15/eval/mid_train/step_5000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15/eval/post_train/step_5000.json"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15",
@@ -135,7 +145,7 @@
           "training_data_mix": "demon_attack_train__bridge",
           "dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps_ghost15",
             "source_subdir": null
           },
@@ -181,7 +191,7 @@
           "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
           "training_recipe_dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps_ghost15",
             "source_subdir": null
           }
@@ -190,7 +200,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps_ghost15"
         }
       ]
@@ -202,9 +212,14 @@
         "path": "openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2"
       },
       "checkpoint": [],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2"
+      },
       "evaluation": [],
       "asset_state": "metadata_only",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2",
@@ -268,7 +283,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2"
         }
       ]
@@ -282,44 +297,49 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
-          "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_exp2/checkpoints/steps_5000_model.safetensors"
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+          "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15/checkpoints/steps_5000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_exp2"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_exp2/eval/mid_train/step_1000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_exp2/eval/mid_train/step_2000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_exp2/eval/mid_train/step_3000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_exp2/eval/mid_train/step_4000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_exp2/eval/mid_train/step_5000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_exp2/eval/post_train/step_5000.json"
         }
       ],
-      "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "asset_state": "shared_checkpoint",
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_exp2",
@@ -335,7 +355,7 @@
           "training_data_mix": "demon_attack_train__bridge",
           "dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps_ghost15",
             "source_subdir": null
           },
@@ -381,7 +401,7 @@
           "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
           "training_recipe_dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps_ghost15",
             "source_subdir": null
           }
@@ -390,7 +410,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps_ghost15"
         }
       ]
@@ -404,44 +424,49 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
-          "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_seed1/checkpoints/steps_5000_model.safetensors"
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+          "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15/checkpoints/steps_5000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_seed1"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_seed1/eval/mid_train/step_1000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_seed1/eval/mid_train/step_2000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_seed1/eval/mid_train/step_3000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_seed1/eval/mid_train/step_4000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_seed1/eval/mid_train/step_5000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_seed1/eval/post_train/step_5000.json"
         }
       ],
-      "asset_state": "complete_bundle",
-      "paper_source_confirmed": false,
+      "asset_state": "shared_checkpoint",
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_seed1",
@@ -457,7 +482,7 @@
           "training_data_mix": "demon_attack_train__bridge",
           "dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps_ghost15",
             "source_subdir": null
           },
@@ -503,7 +528,7 @@
           "interpretation": "The paper experiment type is known; an exact plotted checkpoint binding is recorded only when its immutable source manifest establishes it.",
           "training_recipe_dataset_reference": {
             "repo": "latency-sensitive-bench/benchmark-datasets",
-            "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+            "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
             "path": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps_ghost15",
             "source_subdir": null
           }
@@ -512,7 +537,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps_ghost15"
         }
       ]
@@ -526,29 +551,34 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/demon-attack/small-policy/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/train/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/checkpoint_p0/checkpoint_000048880_25026560.pth"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/demon-attack/small-policy/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/train/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/episode_metrics.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/train/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/git.diff"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/train/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/sf_log.txt"
         }
       ],
       "asset_state": "teacher_bundle",
-      "paper_source_confirmed": false,
+      "paper_checkpoint_binding_confirmed": false,
       "original_training_bindings": [
         {
           "run": "memory/rollout_teachers/demon_attack",
@@ -576,7 +606,7 @@
       "source_prefix": "demon_attack_fixed_latency_6_200ep_7k2steps",
       "prefix": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps",
       "config_name": "demon_attack_fixed_latency_6_200ep_7k2steps",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -587,7 +617,7 @@
       "source_prefix": "demon_attack_fixed_latency_0_200ep_7k2steps",
       "prefix": "zero-latency/demon-attack/demon_attack_fixed_latency_0_200ep_7k2steps",
       "config_name": "demon_attack_fixed_latency_0_200ep_7k2steps",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -598,100 +628,8 @@
       "source_prefix": "training/demon_attack_fixed_latency_6_200ep_7k2steps_ghost15",
       "prefix": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps_ghost15",
       "config_name": "demon_attack_fixed_latency_6_200ep_7k2steps_ghost15",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     }
-  ],
-  "paper_rows": [
-    {
-      "task_key": "demon_attack",
-      "method_key": "single_baseline",
-      "checkpoint_step": "4000",
-      "source_history_step": "4000",
-      "latency_raw_frames": "6",
-      "mean_reward": "5671.75",
-      "std_reward": "633.3815496997051",
-      "mean_length": "14368.08",
-      "std_length": "181.5048032422283",
-      "episode_count": "100",
-      "reward_ci95_lo": "5545.44022353271",
-      "reward_ci95_hi": "5798.05977646729",
-      "length_ci95_lo": "14331.884074911663",
-      "length_ci95_hi": "14404.275925088337",
-      "status": "complete",
-      "source_run_id": "nsb0fpng"
-    },
-    {
-      "task_key": "demon_attack",
-      "method_key": "plain_multi",
-      "checkpoint_step": "4000",
-      "source_history_step": "4000",
-      "latency_raw_frames": "6",
-      "mean_reward": "6474.55",
-      "std_reward": "590.6941234683142",
-      "mean_length": "14342.57",
-      "std_length": "163.35894557691051",
-      "episode_count": "100",
-      "reward_ci95_lo": "6356.753005864003",
-      "reward_ci95_hi": "6592.346994135997",
-      "length_ci95_lo": "14309.992745563815",
-      "length_ci95_hi": "14375.147254436184",
-      "status": "complete",
-      "source_run_id": "dariw0bn"
-    },
-    {
-      "task_key": "demon_attack",
-      "method_key": "stitch",
-      "checkpoint_step": "4000",
-      "source_history_step": "4000",
-      "latency_raw_frames": "6",
-      "mean_reward": "5783.2",
-      "std_reward": "1065.7545496032376",
-      "mean_length": "14400.0",
-      "std_length": "0.0",
-      "episode_count": "100",
-      "reward_ci95_lo": "5570.6658347828325",
-      "reward_ci95_hi": "5995.734165217167",
-      "length_ci95_lo": "14400.0",
-      "length_ci95_hi": "14400.0",
-      "status": "complete",
-      "source_run_id": "aflyy974"
-    },
-    {
-      "task_key": "demon_attack",
-      "method_key": "kv_memory_flex",
-      "checkpoint_step": "4000",
-      "source_history_step": "4000",
-      "latency_raw_frames": "6",
-      "mean_reward": "6483.5",
-      "std_reward": "550.8922308401163",
-      "mean_length": "14370.24",
-      "std_length": "101.39468625130216",
-      "episode_count": "100",
-      "reward_ci95_lo": "6373.640351312772",
-      "reward_ci95_hi": "6593.359648687228",
-      "length_ci95_lo": "14350.019739145471",
-      "length_ci95_hi": "14390.460260854528",
-      "status": "complete",
-      "source_run_id": "6anuta9n"
-    },
-    {
-      "task_key": "demon_attack",
-      "method_key": "ghost_trail",
-      "checkpoint_step": "4000",
-      "source_history_step": "4000",
-      "latency_raw_frames": "6",
-      "mean_reward": "5688.7",
-      "std_reward": "1568.0109406506065",
-      "mean_length": "13459.91",
-      "std_length": "2335.200291602414",
-      "episode_count": "100",
-      "reward_ci95_lo": "5376.005208833846",
-      "reward_ci95_hi": "6001.3947911661535",
-      "length_ci95_lo": "12994.221305754807",
-      "length_ci95_hi": "13925.598694245193",
-      "status": "complete",
-      "source_run_id": "hf:latency-sensitive-bench/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2_exp3/eval/post_train/step_4000.json"
-    }
   ]
 }

experiments/visual-history/archive/flappy/README.md

@@ -2,16 +2,16 @@
 
 | Training run | Input and latency convention | Retained checkpoint | State |
 |---|---|---|---|
-| [flappy_fix_latency_0_200ep_7k2steps_single_baseline](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep_7k2steps_single_baseline) | image_mode=single, num_obs_frames=1, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
-| [flappy_fix_latency_2_200ep_7k2steps_kv_memory](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_7k2steps_kv_memory) | image_mode=multiframe, num_obs_frames=8, target latency unit not recorded in source config | steps_4000_pytorch_model.pt | complete_bundle |
-| [flappy_fix_latency_2_200ep_7k2steps_kv_memory_flex](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_7k2steps_kv_memory_flex) | image_mode=multiframe, num_obs_frames=8, target latency unit not recorded in source config | steps_4000_model.safetensors | complete_bundle |
-| [flappy_fix_latency_2_200ep_7k2steps_plain_multi](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_7k2steps_plain_multi) | image_mode=multiframe, num_obs_frames=4, target latency unit not recorded in source config | steps_4000_model.safetensors | complete_bundle |
-| [flappy_fix_latency_2_200ep_7k2steps_single_baseline](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_7k2steps_single_baseline) | image_mode=single, num_obs_frames=1, target latency unit not recorded in source config | steps_4000_model.safetensors | complete_bundle |
-| [flappy_fix_latency_2_200ep_7k2steps_stitch](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_7k2steps_stitch) | image_mode=stitch, num_obs_frames=4, target latency unit not recorded in source config | steps_4000_model.safetensors | complete_bundle |
-| [openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_model.safetensors | complete_bundle |
-| [openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_model.safetensors | complete_bundle |
-| [openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_model.safetensors | complete_bundle |
-| [memory_rollout_teacher:flappy:teacher:fixed_l3:fs4:obs30:stride1:seed0](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/small-policy/memory_rollout_teacher:flappy:teacher:fixed_l3:fs4:obs30:stride1:seed0) | Original teacher configuration | checkpoint_000048840_25034752.pth | teacher_bundle |
+| [flappy_fix_latency_0_200ep_7k2steps_single_baseline](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep_7k2steps_single_baseline) | image_mode=single, num_obs_frames=1, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
+| [flappy_fix_latency_2_200ep_7k2steps_kv_memory](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_7k2steps_kv_memory) | image_mode=multiframe, num_obs_frames=8, target latency unit not recorded in source config | steps_4000_pytorch_model.pt | complete_bundle |
+| [flappy_fix_latency_2_200ep_7k2steps_kv_memory_flex](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_7k2steps_kv_memory_flex) | image_mode=multiframe, num_obs_frames=8, target latency unit not recorded in source config | steps_4000_model.safetensors | complete_bundle |
+| [flappy_fix_latency_2_200ep_7k2steps_plain_multi](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_7k2steps_plain_multi) | image_mode=multiframe, num_obs_frames=4, target latency unit not recorded in source config | steps_4000_model.safetensors | complete_bundle |
+| [flappy_fix_latency_2_200ep_7k2steps_single_baseline](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_7k2steps_single_baseline) | image_mode=single, num_obs_frames=1, target latency unit not recorded in source config | steps_4000_model.safetensors | complete_bundle |
+| [flappy_fix_latency_2_200ep_7k2steps_stitch](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_7k2steps_stitch) | image_mode=stitch, num_obs_frames=4, target latency unit not recorded in source config | steps_4000_model.safetensors | complete_bundle |
+| [openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_model.safetensors | complete_bundle |
+| [openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_model.safetensors | shared_checkpoint |
+| [openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_model.safetensors | shared_checkpoint |
+| [memory_rollout_teacher:flappy:teacher:fixed_l3:fs4:obs30:stride1:seed0](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/small-policy/memory_rollout_teacher:flappy:teacher:fixed_l3:fs4:obs30:stride1:seed0) | Original teacher configuration | checkpoint_000048840_25034752.pth | teacher_bundle |
 
 ## Asset references
 

experiments/visual-history/deadly-corridor/README.md

@@ -2,11 +2,11 @@
 
 | Training run | Input and latency convention | Retained checkpoint | State |
 |---|---|---|---|
-| [deadly_corridor_fix_latency_6_1000ep_7k2steps_kv_memory_flex](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_6_1000ep_7k2steps_kv_memory_flex) | image_mode=multiframe, num_obs_frames=8, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
-| [deadly_corridor_fix_latency_6_1000ep_7k2steps_plain_multi](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_6_1000ep_7k2steps_plain_multi) | image_mode=multiframe, num_obs_frames=4, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
-| [deadly_corridor_fix_latency_6_1000ep_7k2steps_single_baseline](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_6_1000ep_7k2steps_single_baseline) | image_mode=single, num_obs_frames=1, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
-| [deadly_corridor_fix_latency_6_1000ep_7k2steps_stitch](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_6_1000ep_7k2steps_stitch) | image_mode=stitch, num_obs_frames=4, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
-| [wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/deadly-corridor/vla/starvla-wanoft-h8/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_1000_pytorch_model.pt | complete_bundle |
+| [deadly_corridor_fix_latency_6_1000ep_7k2steps_kv_memory_flex](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_6_1000ep_7k2steps_kv_memory_flex) | image_mode=multiframe, num_obs_frames=8, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
+| [deadly_corridor_fix_latency_6_1000ep_7k2steps_plain_multi](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_6_1000ep_7k2steps_plain_multi) | image_mode=multiframe, num_obs_frames=4, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
+| [deadly_corridor_fix_latency_6_1000ep_7k2steps_single_baseline](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_6_1000ep_7k2steps_single_baseline) | image_mode=single, num_obs_frames=1, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
+| [deadly_corridor_fix_latency_6_1000ep_7k2steps_stitch](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_6_1000ep_7k2steps_stitch) | image_mode=stitch, num_obs_frames=4, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
+| [wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/deadly-corridor/vla/starvla-wanoft-h8/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_1000_pytorch_model.pt | complete_bundle |
 
 ## Evaluation settings
 

experiments/visual-history/deadly-corridor/manifest.json

@@ -12,39 +12,44 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_6_1000ep_7k2steps_kv_memory_flex/checkpoints/steps_4000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_6_1000ep_7k2steps_kv_memory_flex"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_kv_memory_flex/eval/post_train/step_4000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_kv_memory_flex/hydra/.hydra/hydra.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_kv_memory_flex/hydra/.hydra/overrides.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_kv_memory_flex/hydra/train_starvla_hydra.log"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_kv_memory_flex/summary.jsonl"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "run": "deadly_corridor_fix_latency_6_1000ep_7k2steps_kv_memory_flex",
@@ -72,7 +77,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/deadly-corridor/deadly_corridor_fixed_latency_6_1000ep_7k2steps",
           "config_name": "deadly_corridor_fixed_latency_6_1000ep_7k2steps"
         }
@@ -87,39 +92,44 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_6_1000ep_7k2steps_plain_multi/checkpoints/steps_4000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_6_1000ep_7k2steps_plain_multi"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_plain_multi/eval/post_train/step_4000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_plain_multi/hydra/.hydra/hydra.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_plain_multi/hydra/.hydra/overrides.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_plain_multi/hydra/train_starvla_hydra.log"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_plain_multi/summary.jsonl"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "run": "deadly_corridor_fix_latency_6_1000ep_7k2steps_plain_multi",
@@ -147,7 +157,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/deadly-corridor/deadly_corridor_fixed_latency_6_1000ep_7k2steps",
           "config_name": "deadly_corridor_fixed_latency_6_1000ep_7k2steps"
         }
@@ -162,49 +172,54 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_6_1000ep_7k2steps_single_baseline/checkpoints/steps_4000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_6_1000ep_7k2steps_single_baseline"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_single_baseline/eval/post_train/step_2000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_single_baseline/eval/post_train/step_4000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_single_baseline/eval_inputs/deadly_corridor_fixed_latency_6.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_single_baseline/hydra/.hydra/hydra.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_single_baseline/hydra/.hydra/overrides.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_single_baseline/hydra/train_starvla_hydra.log"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_single_baseline/summary.jsonl"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "run": "deadly_corridor_fix_latency_6_1000ep_7k2steps_single_baseline",
@@ -232,7 +247,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/deadly-corridor/deadly_corridor_fixed_latency_6_1000ep_7k2steps",
           "config_name": "deadly_corridor_fixed_latency_6_1000ep_7k2steps"
         }
@@ -247,39 +262,44 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_6_1000ep_7k2steps_stitch/checkpoints/steps_4000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/deadly_corridor_fix_latency_6_1000ep_7k2steps_stitch"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_stitch/eval/post_train/step_4000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_stitch/hydra/.hydra/hydra.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_stitch/hydra/.hydra/overrides.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_stitch/hydra/train_starvla_hydra.log"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/deadly_corridor_fix_latency_6_1000ep_7k2steps_stitch/summary.jsonl"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "run": "deadly_corridor_fix_latency_6_1000ep_7k2steps_stitch",
@@ -307,7 +327,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/deadly-corridor/deadly_corridor_fixed_latency_6_1000ep_7k2steps",
           "config_name": "deadly_corridor_fixed_latency_6_1000ep_7k2steps"
         }
@@ -322,29 +342,34 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/deadly-corridor/vla/starvla-wanoft-h8/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce/checkpoints/steps_1000_pytorch_model.pt"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/deadly-corridor/vla/starvla-wanoft-h8/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce/eval/fixed_6/eval_latency_6/episode_metrics.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce/eval/fixed_6/queue_eval_latency_6.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/deadly-corridor/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce/eval/fixed_6/queue_eval_results.jsonl"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "run": "wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce",
@@ -372,7 +397,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/deadly-corridor/deadly_corridor_fixed_latency_6_1000ep_7k2steps",
           "config_name": "deadly_corridor_fixed_latency_6_1000ep_7k2steps"
         }
@@ -388,7 +413,7 @@
       "source_prefix": "deadly_corridor_fixed_latency_6_1000ep_7k2steps",
       "prefix": "latency-aware/deadly-corridor/deadly_corridor_fixed_latency_6_1000ep_7k2steps",
       "config_name": "deadly_corridor_fixed_latency_6_1000ep_7k2steps",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -399,7 +424,7 @@
       "source_prefix": "deadly_corridor_fixed_latency_0_1000ep_7k2steps",
       "prefix": "zero-latency/deadly-corridor/deadly_corridor_fixed_latency_0_1000ep_7k2steps",
       "config_name": "deadly_corridor_fixed_latency_0_1000ep_7k2steps",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     }
   ],

experiments/visual-history/demon-attack/README.md

@@ -2,12 +2,12 @@
 
 | Training run | Input and latency convention | Retained checkpoint | State |
 |---|---|---|---|
-| [demon_attack_fix_latency_6_200ep_7k2steps_kv_memory_flex](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep_7k2steps_kv_memory_flex) | image_mode=multiframe, num_obs_frames=8, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
-| [demon_attack_fix_latency_6_200ep_7k2steps_plain_multi](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep_7k2steps_plain_multi) | image_mode=multiframe, num_obs_frames=4, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
-| [demon_attack_fix_latency_6_200ep_7k2steps_single_baseline](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep_7k2steps_single_baseline) | image_mode=single, num_obs_frames=1, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
-| [demon_attack_fix_latency_6_200ep_7k2steps_stitch](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep_7k2steps_stitch) | image_mode=stitch, num_obs_frames=4, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
-| [openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2_exp3](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2_exp3) | image_mode=single, num_obs_frames=1, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
-| [wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_pytorch_model.pt | complete_bundle |
+| [demon_attack_fix_latency_6_200ep_7k2steps_kv_memory_flex](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep_7k2steps_kv_memory_flex) | image_mode=multiframe, num_obs_frames=8, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
+| [demon_attack_fix_latency_6_200ep_7k2steps_plain_multi](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep_7k2steps_plain_multi) | image_mode=multiframe, num_obs_frames=4, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
+| [demon_attack_fix_latency_6_200ep_7k2steps_single_baseline](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep_7k2steps_single_baseline) | image_mode=single, num_obs_frames=1, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
+| [demon_attack_fix_latency_6_200ep_7k2steps_stitch](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep_7k2steps_stitch) | image_mode=stitch, num_obs_frames=4, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
+| [openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2_exp3](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2_exp3) | image_mode=single, num_obs_frames=1, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
+| [wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_5000_pytorch_model.pt | complete_bundle |
 
 ## Evaluation settings
 

experiments/visual-history/demon-attack/manifest.json

@@ -12,44 +12,49 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep_7k2steps_kv_memory_flex/checkpoints/steps_4000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep_7k2steps_kv_memory_flex"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_kv_memory_flex/eval/post_train/step_4000.cap3600.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_kv_memory_flex/eval/post_train/step_4000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_kv_memory_flex/hydra/.hydra/hydra.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_kv_memory_flex/hydra/.hydra/overrides.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_kv_memory_flex/hydra/train_starvla_hydra.log"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_kv_memory_flex/summary.jsonl"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "run": "demon_attack_fix_latency_6_200ep_7k2steps_kv_memory_flex",
@@ -77,7 +82,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps",
           "config_name": "demon_attack_fixed_latency_6_200ep_7k2steps"
         }
@@ -92,49 +97,54 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep_7k2steps_plain_multi/checkpoints/steps_4000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep_7k2steps_plain_multi"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_plain_multi/eval/post_train/step_4000.cap3600.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_plain_multi/eval/post_train/step_4000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_plain_multi/eval/post_train/step_4000.pre_framealign.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_plain_multi/hydra/.hydra/hydra.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_plain_multi/hydra/.hydra/overrides.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_plain_multi/hydra/train_starvla_hydra.log"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_plain_multi/summary.jsonl"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "run": "demon_attack_fix_latency_6_200ep_7k2steps_plain_multi",
@@ -162,7 +172,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps",
           "config_name": "demon_attack_fixed_latency_6_200ep_7k2steps"
         }
@@ -177,44 +187,49 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep_7k2steps_single_baseline/checkpoints/steps_4000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep_7k2steps_single_baseline"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_single_baseline/eval/post_train/step_4000.cap3600.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_single_baseline/eval/post_train/step_4000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_single_baseline/hydra/.hydra/hydra.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_single_baseline/hydra/.hydra/overrides.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_single_baseline/hydra/train_starvla_hydra.log"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_single_baseline/summary.jsonl"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "run": "demon_attack_fix_latency_6_200ep_7k2steps_single_baseline",
@@ -242,7 +257,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps",
           "config_name": "demon_attack_fixed_latency_6_200ep_7k2steps"
         }
@@ -257,49 +272,54 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep_7k2steps_stitch/checkpoints/steps_4000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep_7k2steps_stitch"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_stitch/eval/post_train/step_4000.cap3600.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_stitch/eval/post_train/step_4000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_stitch/eval/post_train/step_4000.pre_framealign.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_stitch/hydra/.hydra/hydra.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_stitch/hydra/.hydra/overrides.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_stitch/hydra/train_starvla_hydra.log"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/demon_attack_fix_latency_6_200ep_7k2steps_stitch/summary.jsonl"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "run": "demon_attack_fix_latency_6_200ep_7k2steps_stitch",
@@ -327,7 +347,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps",
           "config_name": "demon_attack_fixed_latency_6_200ep_7k2steps"
         }
@@ -342,54 +362,59 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2_exp3/checkpoints/steps_4000_model.safetensors"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2_exp3"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2_exp3/eval/mid_train/step_1000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2_exp3/eval/mid_train/step_2000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2_exp3/eval/mid_train/step_3000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2_exp3/eval/mid_train/step_4000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2_exp3/eval/post_train/step_4000.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2_exp3/eval/post_train/step_4000.max3600.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2_exp3/eval/post_train/step_4000.part0of2.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2_exp3/eval/post_train/step_4000.part1of2.json"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "source_repo": "latency-sensitive-bench/openvla_bridge_demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2_exp3",
@@ -460,7 +485,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps_ghost15_2"
         }
       ]
@@ -474,49 +499,54 @@
       "checkpoint": [
         {
           "repo": "latency-sensitive-bench/benchmark-models",
-          "revision": "a4099f759f71ffa72d967be3695c8e054cfc377c",
+          "revision": "403deb18146f6780ac28536f78167b21aa18898b",
           "path": "latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/checkpoints/steps_5000_pytorch_model.pt"
         }
       ],
+      "metadata": {
+        "repo": "latency-sensitive-bench/benchmark-models",
+        "revision": "403deb18146f6780ac28536f78167b21aa18898b",
+        "path": "latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce"
+      },
       "evaluation": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/eval/fixed_6/eval_latency_6/episode_metrics.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/eval/fixed_6/queue_eval_latency_6.json"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/eval/fixed_6/queue_eval_results.jsonl"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/hydra/.hydra/hydra.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/hydra/.hydra/overrides.yaml"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/hydra/train_starvla_hydra.log"
         },
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "evaluations/demon-attack/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/summary.jsonl"
         }
       ],
       "asset_state": "complete_bundle",
-      "paper_source_confirmed": true,
+      "paper_checkpoint_binding_confirmed": true,
       "original_training_bindings": [
         {
           "run": "wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce",
@@ -544,7 +574,7 @@
       "training_data": [
         {
           "repo": "latency-sensitive-bench/benchmark-datasets",
-          "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+          "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
           "path": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps",
           "config_name": "demon_attack_fixed_latency_6_200ep_7k2steps"
         }
@@ -560,7 +590,7 @@
       "source_prefix": "demon_attack_fixed_latency_6_200ep_7k2steps",
       "prefix": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps",
       "config_name": "demon_attack_fixed_latency_6_200ep_7k2steps",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -571,7 +601,7 @@
       "source_prefix": "demon_attack_fixed_latency_0_200ep_7k2steps",
       "prefix": "zero-latency/demon-attack/demon_attack_fixed_latency_0_200ep_7k2steps",
       "config_name": "demon_attack_fixed_latency_0_200ep_7k2steps",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     },
     {
@@ -582,7 +612,7 @@
       "source_prefix": "training/demon_attack_fixed_latency_6_200ep_7k2steps_ghost15",
       "prefix": "latency-aware/demon-attack/demon_attack_fixed_latency_6_200ep_7k2steps_ghost15",
       "config_name": "demon_attack_fixed_latency_6_200ep_7k2steps_ghost15",
-      "revision": "15d17287246bd19940c837d6c8a5327564f969c5",
+      "revision": "27d2fe8badf83302809f5570e3118bd987f15600",
       "repo": "latency-sensitive-bench/benchmark-datasets"
     }
   ],

experiments/visual-history/flappy/README.md

@@ -2,12 +2,12 @@
 
 | Training run | Input and latency convention | Retained checkpoint | State |
 |---|---|---|---|
-| [flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex) | image_mode=multiframe, num_obs_frames=8, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
-| [flappy_fix_latency_3_200ep_7k2steps_plain_multi](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep_7k2steps_plain_multi) | image_mode=multiframe, num_obs_frames=4, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
-| [flappy_fix_latency_3_200ep_7k2steps_single_baseline](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep_7k2steps_single_baseline) | image_mode=single, num_obs_frames=1, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
-| [flappy_fix_latency_3_200ep_7k2steps_stitch](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep_7k2steps_stitch) | image_mode=stitch, num_obs_frames=4, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
-| [openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_4000_model.safetensors | complete_bundle |
-| [wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/a4099f759f71ffa72d967be3695c8e054cfc377c/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_2000_pytorch_model.pt | complete_bundle |
+| [flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex) | image_mode=multiframe, num_obs_frames=8, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
+| [flappy_fix_latency_3_200ep_7k2steps_plain_multi](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep_7k2steps_plain_multi) | image_mode=multiframe, num_obs_frames=4, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
+| [flappy_fix_latency_3_200ep_7k2steps_single_baseline](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep_7k2steps_single_baseline) | image_mode=single, num_obs_frames=1, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
+| [flappy_fix_latency_3_200ep_7k2steps_stitch](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep_7k2steps_stitch) | image_mode=stitch, num_obs_frames=4, target latency unit=raw_frames | steps_4000_model.safetensors | complete_bundle |
+| [openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_4000_model.safetensors | complete_bundle |
+| [wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce](https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/403deb18146f6780ac28536f78167b21aa18898b/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce) | image_mode=single, num_obs_frames=1, target latency unit=observation_steps | steps_2000_pytorch_model.pt | complete_bundle |
 
 ## Evaluation settings