User simulation leaderboard: results

Built on the Hugging Face leaderboard template (demo-leaderboard-backend). Five behavioral
benchmarks, Recall@10, systems typed trained/prompted/baseline with a Verification column.
Numbers are exported from Jean's evaluation repository, not typed.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

.gitattributes

@@ -0,0 +1,55 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.lz4 filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+# Audio files - uncompressed
+*.pcm filter=lfs diff=lfs merge=lfs -text
+*.sam filter=lfs diff=lfs merge=lfs -text
+*.raw filter=lfs diff=lfs merge=lfs -text
+# Audio files - compressed
+*.aac filter=lfs diff=lfs merge=lfs -text
+*.flac filter=lfs diff=lfs merge=lfs -text
+*.mp3 filter=lfs diff=lfs merge=lfs -text
+*.ogg filter=lfs diff=lfs merge=lfs -text
+*.wav filter=lfs diff=lfs merge=lfs -text
+# Image files - uncompressed
+*.bmp filter=lfs diff=lfs merge=lfs -text
+*.gif filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.tiff filter=lfs diff=lfs merge=lfs -text
+# Image files - compressed
+*.jpg filter=lfs diff=lfs merge=lfs -text
+*.jpeg filter=lfs diff=lfs merge=lfs -text
+*.webp filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,22 @@
+---
+license: apache-2.0
+tags:
+  - leaderboard
+---
+
+# User simulation leaderboard: results
+
+One file per system per release, under `<org>/<system>/results_<timestamp>.json`.
+
+`results` holds the score the leaderboard displays (Recall@10 per benchmark). `record` holds
+everything that makes the row citable and is not shown in the grid: the person-clustered
+bootstrap interval, n, NDCG@10, the popularity and stranger controls, the evaluation cell, and
+the run that produced it (experiment id, commit, date, seed).
+
+These files are generated, never edited: `tools/leaderboard_export.py` in Jean's evaluation
+repository writes the site's `leaderboard.json` from each run's own `results.json`, and
+`convert.py` in the leaderboard repository turns that into the files here. A rerun adds a file;
+no file is rewritten.
+
+Leaderboard: https://jeantechnologies.com/evals · Methodology:
+https://jeantechnologies.com/docs/leaderboard-methodology

baselines/popularity/results_2026-09-13T18-57-44.json

@@ -0,0 +1,133 @@
+{
+  "config": {
+    "model_dtype": "float32",
+    "model_name": "baselines/popularity",
+    "model_sha": "682e914",
+    "display_name": "Popularity",
+    "architecture": "The most frequent training items, in the same order for every person. The floor every row must clear.",
+    "url": "https://jeantechnologies.com/evals"
+  },
+  "results": {
+    "amazon2014-beauty": {
+      "recall@10": 0.012
+    },
+    "steam-2018": {
+      "recall@10": 0.052
+    },
+    "amazon2014-five-category-transfer": {
+      "recall@10": 0.0098
+    },
+    "amazon2014-taste-transfer": {
+      "recall@10": 0.0164
+    },
+    "douban-tridomain-transfer": {
+      "recall@10": 0.0257
+    }
+  },
+  "record": {
+    "amazon2014-beauty": {
+      "recall@10": 0.012,
+      "interval": null,
+      "n": 22363,
+      "ndcg@10": 0.0056,
+      "controls": {},
+      "cell": "individual, full observed history, next item in domain",
+      "run": {
+        "experiment": "020-calibrate-beauty",
+        "commit": "f245c74",
+        "date": "2026-09-04",
+        "seed": 0
+      },
+      "cost": {
+        "fit_seconds": 1.1,
+        "usd_per_1k": null
+      }
+    },
+    "steam-2018": {
+      "recall@10": 0.052,
+      "interval": [
+        0.0491,
+        0.0551
+      ],
+      "n": 20000,
+      "ndcg@10": 0.0267,
+      "controls": {},
+      "cell": "individual, full observed history, next item in domain",
+      "run": {
+        "experiment": "033-trait-floor-steam",
+        "commit": "8a5d439",
+        "date": "2026-09-06",
+        "seed": 0
+      },
+      "cost": {
+        "fit_seconds": 2.1,
+        "usd_per_1k": null
+      }
+    },
+    "amazon2014-five-category-transfer": {
+      "recall@10": 0.0098,
+      "interval": [
+        0.0078,
+        0.012
+      ],
+      "n": 8348,
+      "ndcg@10": 0.0068,
+      "controls": {},
+      "cell": "individual, history from four other categories, next item in an unseen category",
+      "run": {
+        "experiment": "016-transfer-holdout",
+        "commit": "8637e67",
+        "date": "2026-09-04",
+        "seed": 0
+      },
+      "cost": {
+        "fit_seconds": 1.5,
+        "usd_per_1k": null
+      }
+    },
+    "amazon2014-taste-transfer": {
+      "recall@10": 0.0164,
+      "interval": [
+        0.0115,
+        0.0216
+      ],
+      "n": 2686,
+      "ndcg@10": 0.0082,
+      "controls": {},
+      "cell": "individual, history from other taste categories, next item in an unseen category",
+      "run": {
+        "experiment": "034b-trait-transfer-dedup",
+        "commit": "7b20023",
+        "date": "2026-09-06",
+        "seed": 0
+      },
+      "cost": {
+        "fit_seconds": 1.9,
+        "usd_per_1k": null
+      }
+    },
+    "douban-tridomain-transfer": {
+      "recall@10": 0.0257,
+      "interval": [
+        0.023,
+        0.0285
+      ],
+      "n": 12139,
+      "ndcg@10": 0.0132,
+      "controls": {},
+      "cell": "individual, history from two other domains, next item in an unseen domain, trained on the full panel",
+      "run": {
+        "experiment": "045-transfer-douban",
+        "commit": "4d4b218",
+        "date": "2026-09-08",
+        "seed": 0
+      },
+      "cost": {
+        "fit_seconds": 2.0,
+        "usd_per_1k": null
+      }
+    }
+  },
+  "leaderboard_release": "2026-09-12",
+  "eval_code_commit": "682e914"
+}

jean/sasrec-trained-from-scratch/results_2026-09-13T18-57-44.json

@@ -0,0 +1,146 @@
+{
+  "config": {
+    "model_dtype": "float32",
+    "model_name": "jean/sasrec-trained-from-scratch",
+    "model_sha": "682e914",
+    "display_name": "SASRec, trained from scratch",
+    "architecture": "The basis of jean-rec-1. Two layers, width 64, twenty-item window, full-softmax loss, trained on the snapshot's own training split. training_method: trained_from_scratch.",
+    "url": "https://jeantechnologies.com/evals"
+  },
+  "results": {
+    "amazon2014-beauty": {
+      "recall@10": 0.094
+    },
+    "steam-2018": {
+      "recall@10": 0.1115
+    },
+    "amazon2014-five-category-transfer": {
+      "recall@10": 0.0416
+    },
+    "amazon2014-taste-transfer": {
+      "recall@10": 0.0331
+    },
+    "douban-tridomain-transfer": {
+      "recall@10": 0.0523
+    }
+  },
+  "record": {
+    "amazon2014-beauty": {
+      "recall@10": 0.094,
+      "interval": null,
+      "n": 22363,
+      "ndcg@10": 0.0562,
+      "controls": {
+        "popularity": 0.012
+      },
+      "cell": "individual, full observed history, next item in domain",
+      "run": {
+        "experiment": "020-calibrate-beauty",
+        "commit": "f245c74",
+        "date": "2026-09-04",
+        "seed": 0
+      },
+      "cost": {
+        "fit_seconds": 136.7,
+        "usd_per_1k": null
+      }
+    },
+    "steam-2018": {
+      "recall@10": 0.1115,
+      "interval": [
+        0.1074,
+        0.1158
+      ],
+      "n": 20000,
+      "ndcg@10": 0.0579,
+      "controls": {
+        "popularity": 0.052
+      },
+      "cell": "individual, full observed history, next item in domain",
+      "run": {
+        "experiment": "033-trait-floor-steam",
+        "commit": "8a5d439",
+        "date": "2026-09-06",
+        "seed": 0
+      },
+      "cost": {
+        "fit_seconds": 1313.1,
+        "usd_per_1k": null
+      }
+    },
+    "amazon2014-five-category-transfer": {
+      "recall@10": 0.0416,
+      "interval": [
+        0.0376,
+        0.0459
+      ],
+      "n": 8348,
+      "ndcg@10": 0.0221,
+      "controls": {
+        "popularity": 0.0098,
+        "stranger": 0.0067
+      },
+      "cell": "individual, history from four other categories, next item in an unseen category",
+      "run": {
+        "experiment": "016-transfer-holdout",
+        "commit": "8637e67",
+        "date": "2026-09-04",
+        "seed": 0
+      },
+      "cost": {
+        "fit_seconds": 1845.7,
+        "usd_per_1k": null
+      }
+    },
+    "amazon2014-taste-transfer": {
+      "recall@10": 0.0331,
+      "interval": [
+        0.0268,
+        0.0398
+      ],
+      "n": 2686,
+      "ndcg@10": 0.0171,
+      "controls": {
+        "popularity": 0.0164,
+        "stranger": 0.0056
+      },
+      "cell": "individual, history from other taste categories, next item in an unseen category",
+      "run": {
+        "experiment": "034b-trait-transfer-dedup",
+        "commit": "7b20023",
+        "date": "2026-09-06",
+        "seed": 0
+      },
+      "cost": {
+        "fit_seconds": 1811.0,
+        "usd_per_1k": null
+      }
+    },
+    "douban-tridomain-transfer": {
+      "recall@10": 0.0523,
+      "interval": [
+        0.0485,
+        0.0562
+      ],
+      "n": 12139,
+      "ndcg@10": 0.0276,
+      "controls": {
+        "popularity": 0.0257,
+        "stranger": 0.0248
+      },
+      "cell": "individual, history from two other domains, next item in an unseen domain, trained on the full panel",
+      "run": {
+        "experiment": "045-transfer-douban",
+        "commit": "4d4b218",
+        "date": "2026-09-08",
+        "seed": 0
+      },
+      "cost": {
+        "fit_seconds": 872.6,
+        "usd_per_1k": null
+      }
+    }
+  },
+  "leaderboard_release": "2026-09-12",
+  "eval_code_commit": "682e914"
+}