Update model card for intermediate checkpoints

020305/README.md

@@ -7,24 +7,28 @@ This release contains the 020305 and 050302 extrapolation checkpoint batches for
 ```text
 020305/
   README.md
-  pretrain/checkpoint-18809/
-  rl/op11_14_step200/
-  rl/op13_20_step200/
-  rl/op15_20_step200/
+  pretrain/checkpoint-{3762,7524,11286,15048,18809}/
+  rl/op11_14_step{20,40,60,80,100,120,140,160,180,200}/
+  rl/op13_20_step{20,40,60,80,100,120,140,160,180,200}/
+  rl/op15_20_step{20,40,60,80,100,120,140,160,180,200}/
+  opd_refkl/sweep/step{20,40,60,80,100,120,140,160,180}/
   opd_refkl/op11_14_step200/
   evals/
 050302/
   README.md
-  pretrain/checkpoint-18586/
-  rl/op11_14_step200/
-  rl/op13_20_step200/
-  rl/op15_20_step200/
+  pretrain/checkpoint-{3718,7436,11154,14872,18586}/
+  rl/op11_14_step{20,40,60,80,100,120,140,160,180,200}/
+  rl/op13_20_step{20,40,60,80,100,120,140,160,180,200}/
+  rl/op15_20_step{20,40,60,80,100,120,140,160,180,200}/
+  opd_refkl/sweep/step{20,40,60,80,100,120,140,160,180}/
   opd_refkl/op13_20_step200/
   evals/
 ```
 
 The default upload includes compact evaluation files: `summary.csv`, `summary.json`, per-model `*_metrics.json`, graph `results.json`, and score reports. Raw generation JSONL files and per-sample graph-result JSONL files are large and are uploaded when the upload helper is run with `--include-generations` and `--include-per-sample-results`.
 
+The non-final pretrain/RL/OPD checkpoints are model-only unless an evaluation artifact path is separately listed below. The 050302 OPD sweep has intermediate evaluation artifacts under `050302/evals/opd_refkl_sweep/`.
+
 ## Run Index
 
 | Model path | Exp name | Script/config | Model name | Time |
@@ -59,6 +63,7 @@ All rows use 1,152,000 ID examples, 1,280,000 OOD examples, and 2,432,000 total
 
 ## Source Pointers
 
+- Model release root: <https://huggingface.co/ZachW/extrapolation_rl>
 - Data: <https://huggingface.co/datasets/Interplay-LM-Reasoning/composition>
 - Context data: <https://huggingface.co/datasets/Interplay-LM-Reasoning/context>
 - Repository documentation: `docs/extrapolation_release_wrapup.md` in the GitHub repository.

050302/README.md

@@ -7,24 +7,28 @@ This release contains the 020305 and 050302 extrapolation checkpoint batches for
 ```text
 020305/
   README.md
-  pretrain/checkpoint-18809/
-  rl/op11_14_step200/
-  rl/op13_20_step200/
-  rl/op15_20_step200/
+  pretrain/checkpoint-{3762,7524,11286,15048,18809}/
+  rl/op11_14_step{20,40,60,80,100,120,140,160,180,200}/
+  rl/op13_20_step{20,40,60,80,100,120,140,160,180,200}/
+  rl/op15_20_step{20,40,60,80,100,120,140,160,180,200}/
+  opd_refkl/sweep/step{20,40,60,80,100,120,140,160,180}/
   opd_refkl/op11_14_step200/
   evals/
 050302/
   README.md
-  pretrain/checkpoint-18586/
-  rl/op11_14_step200/
-  rl/op13_20_step200/
-  rl/op15_20_step200/
+  pretrain/checkpoint-{3718,7436,11154,14872,18586}/
+  rl/op11_14_step{20,40,60,80,100,120,140,160,180,200}/
+  rl/op13_20_step{20,40,60,80,100,120,140,160,180,200}/
+  rl/op15_20_step{20,40,60,80,100,120,140,160,180,200}/
+  opd_refkl/sweep/step{20,40,60,80,100,120,140,160,180}/
   opd_refkl/op13_20_step200/
   evals/
 ```
 
 The default upload includes compact evaluation files: `summary.csv`, `summary.json`, per-model `*_metrics.json`, graph `results.json`, and score reports. Raw generation JSONL files and per-sample graph-result JSONL files are large and are uploaded when the upload helper is run with `--include-generations` and `--include-per-sample-results`.
 
+The non-final pretrain/RL/OPD checkpoints are model-only unless an evaluation artifact path is separately listed below. The 050302 OPD sweep has intermediate evaluation artifacts under `050302/evals/opd_refkl_sweep/`.
+
 ## Run Index
 
 | Model path | Exp name | Script/config | Model name | Time |
@@ -59,6 +63,7 @@ All rows use 1,152,000 ID examples, 1,280,000 OOD examples, and 2,432,000 total
 
 ## Source Pointers
 
+- Model release root: <https://huggingface.co/ZachW/extrapolation_rl>
 - Data: <https://huggingface.co/datasets/Interplay-LM-Reasoning/composition>
 - Context data: <https://huggingface.co/datasets/Interplay-LM-Reasoning/context>
 - Repository documentation: `docs/extrapolation_release_wrapup.md` in the GitHub repository.

README.md

@@ -22,20 +22,21 @@ Project code and documentation: <https://github.com/YichenZW/Interplay-LM-Reason
 ```text
 020305/
   README.md
-  pretrain/checkpoint-18809/
-  rl/op11_14_step200/
-  rl/op13_20_step200/
-  rl/op15_20_step200/
+  pretrain/checkpoint-{3762,7524,11286,15048,18809}/
+  rl/op11_14_step{20,40,60,80,100,120,140,160,180,200}/
+  rl/op13_20_step{20,40,60,80,100,120,140,160,180,200}/
+  rl/op15_20_step{20,40,60,80,100,120,140,160,180,200}/
+  opd_refkl/sweep/step{20,40,60,80,100,120,140,160,180}/
   opd_refkl/op11_14_step200/
   evals/
 050302/
   README.md
-  pretrain/checkpoint-18586/
-  rl/op11_14_step200/
-  rl/op13_20_step200/
-  rl/op15_20_step200/
+  pretrain/checkpoint-{3718,7436,11154,14872,18586}/
+  rl/op11_14_step{20,40,60,80,100,120,140,160,180,200}/
+  rl/op13_20_step{20,40,60,80,100,120,140,160,180,200}/
+  rl/op15_20_step{20,40,60,80,100,120,140,160,180,200}/
+  opd_refkl/sweep/step{20,40,60,80,100,120,140,160,180}/
   opd_refkl/op13_20_step200/
-  opd_refkl/sweep/
   evals/
 ```
 
@@ -82,6 +83,10 @@ Raw generations are uploaded next to the compact evaluation artifacts when avail
 
 The 020305 far-OOD probes are under `020305/evals/far_ood/`. The 050302 OPD intermediate sweep evaluations are under `050302/evals/opd_refkl_sweep/`.
 
+## Non-final Checkpoints
+
+The uploaded non-final pretrain/RL/OPD checkpoints are model-only unless an evaluation directory is listed above. The 050302 OPD sweep has intermediate evaluation artifacts under `050302/evals/opd_refkl_sweep/`; the newly added pretrain, RL, and 020305 OPD non-final checkpoints do not have separate evaluation artifacts in this release.
+
 ## Evaluation Summary
 
 All rows use 1,152,000 ID examples, 1,280,000 OOD examples, and 2,432,000 total examples.