| { |
| "base_model": "allenai/Olmo-3-7B-Think", |
| "checkpoints": [ |
| { |
| "bytes": 13779134611, |
| "description": "Counterfactual sensitivity Behavior Forecaster trained on the main OLMo data surface.", |
| "metrics": { |
| "class_balanced_accuracy": 0.949855857538682, |
| "loss": 0.16984329223632813, |
| "n_negative": 12450, |
| "n_positive": 3870, |
| "pearson_r": 0.9326592336450144, |
| "spearman_r": 0.7270551546398006 |
| }, |
| "name": "counterfactual_sensitivity_main", |
| "path": "counterfactual_sensitivity/main/final_model.pt", |
| "sha256": "f14e7a008a2ada5b024b1156f0afcd1a3f037d8e1cff4788b1255667d800424c" |
| }, |
| { |
| "bytes": 14010245401, |
| "description": "Rerun consistency Behavior Forecaster trained on the main OLMo data surface.", |
| "metrics": { |
| "class_balanced_accuracy": 0.5535714285714286, |
| "loss": 0.02789507378472222, |
| "n_negative": 280, |
| "n_positive": 845, |
| "pearson_r": 0.6291667992532701, |
| "spearman_r": 0.5703879775477773 |
| }, |
| "name": "rerun_consistency_main", |
| "path": "rerun_consistency/main/final_model.pt", |
| "sha256": "6e2e78ed8d9a5ce78605e5db0428e2eab2732c3fe5219d6a547349fe5300d407" |
| }, |
| { |
| "bytes": 13779134419, |
| "description": "Counterfactual sensitivity source model trained without FEVEROUS, for transfer-to-FEVEROUS evaluation.", |
| "metrics": { |
| "class_balanced_accuracy": 0.7221790654816058, |
| "loss": 0.645204319780111, |
| "n_negative": 2165, |
| "n_positive": 990, |
| "pearson_r": 0.5188018919649807, |
| "spearman_r": 0.4940661915609062 |
| }, |
| "name": "counterfactual_sensitivity_transfer_feverous_source", |
| "path": "counterfactual_sensitivity/transfer_to_feverous/source_model/final_model.pt", |
| "sha256": "7f4b343a5befec3e24a57e3c373a3617ec96415e13fce2244ba864f3b96d7b69" |
| }, |
| { |
| "bytes": 13779134803, |
| "description": "Counterfactual sensitivity model adapted on FEVEROUS for cross-dataset transfer.", |
| "metrics": { |
| "class_balanced_accuracy": 0.8298597989129167, |
| "loss": 0.3910848046255943, |
| "n_negative": 2165, |
| "n_positive": 990, |
| "pearson_r": 0.7789758594434558, |
| "spearman_r": 0.6769259317889156 |
| }, |
| "name": "counterfactual_sensitivity_transfer_feverous_adapted", |
| "path": "counterfactual_sensitivity/transfer_to_feverous/adapted_model/final_model.pt", |
| "sha256": "220f2c1829c69e06d2fc7f6975ed7c918ad125fd089e649b872e0a243f0b4132" |
| }, |
| { |
| "bytes": 13779134547, |
| "description": "Counterfactual sensitivity source model trained without TreeCut, for transfer-to-TreeCut evaluation.", |
| "metrics": { |
| "class_balanced_accuracy": 0.622464767911698, |
| "loss": 0.6533952679550439, |
| "n_negative": 695, |
| "n_positive": 730, |
| "pearson_r": 0.5598928526114065, |
| "spearman_r": 0.6126665579472638 |
| }, |
| "name": "counterfactual_sensitivity_transfer_treecut_source", |
| "path": "counterfactual_sensitivity/transfer_to_treecut/source_model/final_model.pt", |
| "sha256": "d0bbb7e1e7ca2ebc6bf5887f28fe982216c1d5f2c1b550be2fa87db1771abf04" |
| }, |
| { |
| "bytes": 13779134803, |
| "description": "Counterfactual sensitivity model adapted on TreeCut for cross-dataset transfer.", |
| "metrics": { |
| "class_balanced_accuracy": 0.961806256147986, |
| "loss": 0.15939776344200443, |
| "n_negative": 695, |
| "n_positive": 730, |
| "pearson_r": 0.9402512548581133, |
| "spearman_r": 0.8465005636349516 |
| }, |
| "name": "counterfactual_sensitivity_transfer_treecut_adapted", |
| "path": "counterfactual_sensitivity/transfer_to_treecut/adapted_model/final_model.pt", |
| "sha256": "94a73e61fac5679ab71e46c1614d4a39eab4bb22e7b5c09837532aec93210a11" |
| } |
| ], |
| "format": "PyTorch checkpoints produced by the Behavior Forecasters OLMo release code path.", |
| "note": "These are inference/evaluation final_model.pt checkpoints, not optimizer/scheduler resume checkpoints.", |
| "repo_id": "Mosh/behavior-forecasters-olmo-checkpoints" |
| } |
|
|
