{ "bench_version": 1, "track": "latent_mixture_transport", "model": "mlp_tiny", "metric_direction": "lower is better", "n_seeds": 8, "baseline": { "best_cfg": { "lr": 0.01 }, "sweep": [ { "per_seed": [ 1.6584807634353638, 1.8023967742919922, 1.948538064956665, 1.7409632205963135, 1.6989234685897827, 2.0602176189422607, 1.787219524383545, 1.790303349494934 ], "mean": 1.8108803480863571, "std": 0.1239063765057777, "config": { "lr": 0.001 }, "signatures": [ { "pred_mean": [ 0.043234892189502716, 0.06797278672456741 ], "pred_std": [ 4.53477668762207, 1.0423473119735718 ], "target_mean": [ 0.10782764106988907, -0.019209526479244232 ], "target_std": [ 4.536637783050537, 1.6189589500427246 ] }, { "pred_mean": [ 0.21873944997787476, -0.13643093407154083 ], "pred_std": [ 4.613937854766846, 1.0739085674285889 ], "target_mean": [ 0.28565511107444763, -0.07340104132890701 ], "target_std": [ 4.524669170379639, 1.6303699016571045 ] }, { "pred_mean": [ 0.24090030789375305, -0.05906757339835167 ], "pred_std": [ 4.5369672775268555, 0.9227220416069031 ], "target_mean": [ 0.37497055530548096, -0.06369480490684509 ], "target_std": [ 4.55037260055542, 1.7904703617095947 ] }, { "pred_mean": [ -0.006168227177113295, -0.1031428799033165 ], "pred_std": [ 4.422000408172607, 0.9418965578079224 ], "target_mean": [ 0.032932933419942856, -0.04044200852513313 ], "target_std": [ 4.487130165100098, 1.6359201669692993 ] }, { "pred_mean": [ 0.41770225763320923, -0.030527109280228615 ], "pred_std": [ 4.762247562408447, 1.0726734399795532 ], "target_mean": [ 0.3325880467891693, 0.018684253096580505 ], "target_std": [ 4.577158451080322, 1.662758231163025 ] }, { "pred_mean": [ 0.3517788052558899, 0.025978371500968933 ], "pred_std": [ 4.633047103881836, 0.9402766823768616 ], "target_mean": [ 0.0659371018409729, 0.149443119764328 ], "target_std": [ 4.564814567565918, 1.66139554977417 ] }, { "pred_mean": [ 0.08992262184619904, -0.0018513768445700407 ], "pred_std": [ 4.472599506378174, 0.9126293063163757 ], "target_mean": [ 0.15029877424240112, -0.006664793472737074 ], "target_std": [ 4.5857696533203125, 1.6789789199829102 ] }, { "pred_mean": [ -0.22141434252262115, 0.0017037141369655728 ], "pred_std": [ 4.363723278045654, 0.849623441696167 ], "target_mean": [ -0.2404521107673645, 0.11181623488664627 ], "target_std": [ 4.359499454498291, 1.6317424774169922 ] } ] }, { "per_seed": [ 1.4588760137557983, 1.3883919715881348, 1.5811883211135864, 1.4916973114013672, 1.4352669715881348, 1.693009614944458, 1.5048573017120361, 1.3684591054916382 ], "mean": 1.4902183264493942, "std": 0.09925487394239246, "config": { "lr": 0.003 }, "signatures": [ { "pred_mean": [ 0.06621045619249344, -0.1021866425871849 ], "pred_std": [ 4.53108549118042, 1.266789436340332 ], "target_mean": [ 0.10782764106988907, -0.019209526479244232 ], "target_std": [ 4.536637783050537, 1.6189589500427246 ] }, { "pred_mean": [ 0.17134340107440948, -0.04508264362812042 ], "pred_std": [ 4.447978973388672, 1.245232343673706 ], "target_mean": [ 0.28565511107444763, -0.07340104132890701 ], "target_std": [ 4.524669170379639, 1.6303699016571045 ] }, { "pred_mean": [ 0.30628085136413574, -0.01341619249433279 ], "pred_std": [ 4.2159857749938965, 1.1691285371780396 ], "target_mean": [ 0.37497055530548096, -0.06369480490684509 ], "target_std": [ 4.55037260055542, 1.7904703617095947 ] }, { "pred_mean": [ -0.03721538558602333, -0.04432598128914833 ], "pred_std": [ 4.213113784790039, 1.017906904220581 ], "target_mean": [ 0.032932933419942856, -0.04044200852513313 ], "target_std": [ 4.487130165100098, 1.6359201669692993 ] }, { "pred_mean": [ 0.5395490527153015, -0.13345450162887573 ], "pred_std": [ 4.522548675537109, 1.2905967235565186 ], "target_mean": [ 0.3325880467891693, 0.018684253096580505 ], "target_std": [ 4.577158451080322, 1.662758231163025 ] }, { "pred_mean": [ 0.2691681385040283, 0.1501285880804062 ], "pred_std": [ 4.459775447845459, 1.1214773654937744 ], "target_mean": [ 0.0659371018409729, 0.149443119764328 ], "target_std": [ 4.564814567565918, 1.66139554977417 ] }, { "pred_mean": [ 0.1690424382686615, -0.08001205325126648 ], "pred_std": [ 4.409627914428711, 1.0867819786071777 ], "target_mean": [ 0.15029877424240112, -0.006664793472737074 ], "target_std": [ 4.5857696533203125, 1.6789789199829102 ] }, { "pred_mean": [ -0.17967349290847778, -0.042357780039310455 ], "pred_std": [ 4.187299728393555, 1.2296702861785889 ], "target_mean": [ -0.2404521107673645, 0.11181623488664627 ], "target_std": [ 4.359499454498291, 1.6317424774169922 ] } ] }, { "per_seed": [ 1.5505059957504272, 1.2503662109375, 1.5140531063079834, 1.4406028985977173, 1.48829984664917, 1.5539131164550781, 1.3900737762451172, 1.3093628883361816 ], "mean": 1.4371472299098969, "std": 0.1051543061191552, "config": { "lr": 0.01 }, "signatures": [ { "pred_mean": [ 0.2688068449497223, -0.1351122260093689 ], "pred_std": [ 4.552779674530029, 1.5561680793762207 ], "target_mean": [ 0.10782764106988907, -0.019209526479244232 ], "target_std": [ 4.536637783050537, 1.6189589500427246 ] }, { "pred_mean": [ 0.24252046644687653, 0.0590505413711071 ], "pred_std": [ 4.520332336425781, 1.3489108085632324 ], "target_mean": [ 0.28565511107444763, -0.07340104132890701 ], "target_std": [ 4.524669170379639, 1.6303699016571045 ] }, { "pred_mean": [ 0.6274702548980713, -0.20266346633434296 ], "pred_std": [ 4.23753023147583, 1.2169533967971802 ], "target_mean": [ 0.37497055530548096, -0.06369480490684509 ], "target_std": [ 4.55037260055542, 1.7904703617095947 ] }, { "pred_mean": [ 0.03630140423774719, -0.12918049097061157 ], "pred_std": [ 4.289375305175781, 1.1092267036437988 ], "target_mean": [ 0.032932933419942856, -0.04044200852513313 ], "target_std": [ 4.487130165100098, 1.6359201669692993 ] }, { "pred_mean": [ 0.5265708565711975, -0.3529617190361023 ], "pred_std": [ 4.702226161956787, 1.4276982545852661 ], "target_mean": [ 0.3325880467891693, 0.018684253096580505 ], "target_std": [ 4.577158451080322, 1.662758231163025 ] }, { "pred_mean": [ 0.15553702414035797, 0.1446627378463745 ], "pred_std": [ 4.38116455078125, 1.2813775539398193 ], "target_mean": [ 0.0659371018409729, 0.149443119764328 ], "target_std": [ 4.564814567565918, 1.66139554977417 ] }, { "pred_mean": [ 0.23208913207054138, -0.04567822068929672 ], "pred_std": [ 4.284600734710693, 1.240485668182373 ], "target_mean": [ 0.15029877424240112, -0.006664793472737074 ], "target_std": [ 4.5857696533203125, 1.6789789199829102 ] }, { "pred_mean": [ -0.31963932514190674, -0.02888385020196438 ], "pred_std": [ 4.275836944580078, 1.2498278617858887 ], "target_mean": [ -0.2404521107673645, 0.11181623488664627 ], "target_std": [ 4.359499454498291, 1.6317424774169922 ] } ] } ], "full": { "per_seed": [ 1.5505059957504272, 1.2503662109375, 1.5140531063079834, 1.4406028985977173, 1.48829984664917, 1.5539131164550781, 1.3900737762451172, 1.3093628883361816 ], "mean": 1.4371472299098969, "std": 0.1051543061191552, "config": { "lr": 0.01 }, "signatures": [ { "pred_mean": [ 0.2688068449497223, -0.1351122260093689 ], "pred_std": [ 4.552779674530029, 1.5561680793762207 ], "target_mean": [ 0.10782764106988907, -0.019209526479244232 ], "target_std": [ 4.536637783050537, 1.6189589500427246 ] }, { "pred_mean": [ 0.24252046644687653, 0.0590505413711071 ], "pred_std": [ 4.520332336425781, 1.3489108085632324 ], "target_mean": [ 0.28565511107444763, -0.07340104132890701 ], "target_std": [ 4.524669170379639, 1.6303699016571045 ] }, { "pred_mean": [ 0.6274702548980713, -0.20266346633434296 ], "pred_std": [ 4.23753023147583, 1.2169533967971802 ], "target_mean": [ 0.37497055530548096, -0.06369480490684509 ], "target_std": [ 4.55037260055542, 1.7904703617095947 ] }, { "pred_mean": [ 0.03630140423774719, -0.12918049097061157 ], "pred_std": [ 4.289375305175781, 1.1092267036437988 ], "target_mean": [ 0.032932933419942856, -0.04044200852513313 ], "target_std": [ 4.487130165100098, 1.6359201669692993 ] }, { "pred_mean": [ 0.5265708565711975, -0.3529617190361023 ], "pred_std": [ 4.702226161956787, 1.4276982545852661 ], "target_mean": [ 0.3325880467891693, 0.018684253096580505 ], "target_std": [ 4.577158451080322, 1.662758231163025 ] }, { "pred_mean": [ 0.15553702414035797, 0.1446627378463745 ], "pred_std": [ 4.38116455078125, 1.2813775539398193 ], "target_mean": [ 0.0659371018409729, 0.149443119764328 ], "target_std": [ 4.564814567565918, 1.66139554977417 ] }, { "pred_mean": [ 0.23208913207054138, -0.04567822068929672 ], "pred_std": [ 4.284600734710693, 1.240485668182373 ], "target_mean": [ 0.15029877424240112, -0.006664793472737074 ], "target_std": [ 4.5857696533203125, 1.6789789199829102 ] }, { "pred_mean": [ -0.31963932514190674, -0.02888385020196438 ], "pred_std": [ 4.275836944580078, 1.2498278617858887 ], "target_mean": [ -0.2404521107673645, 0.11181623488664627 ], "target_std": [ 4.359499454498291, 1.6317424774169922 ] } ] } }, "idea": { "per_seed": [ 0.1718381643295288, 0.13643871247768402, 0.17405973374843597, 0.17571242153644562, 0.14317259192466736, 0.19144359230995178, 0.16002926230430603, 0.1587807536125183 ], "mean": 0.16393440403044224, "std": 0.01691650453456634, "config": { "lr": 0.01, "epsilon": 0.1, "tau": 3.0 }, "signatures": [ { "pred_mean": [ 0.06780313700437546, -0.053052350878715515 ], "pred_std": [ 2.044187545776367, 1.0252753496170044 ], "target_mean": [ 0.04435388743877411, -0.012574340216815472 ], "target_std": [ 2.0263524055480957, 1.0520573854446411 ] }, { "pred_mean": [ 0.09832692891359329, 0.015258383937180042 ], "pred_std": [ 2.0508484840393066, 1.0656163692474365 ], "target_mean": [ 0.13986718654632568, -0.07649480551481247 ], "target_std": [ 2.022428512573242, 1.0598915815353394 ] }, { "pred_mean": [ 0.17367428541183472, -0.08788446336984634 ], "pred_std": [ 1.9044790267944336, 0.9941457509994507 ], "target_mean": [ 0.12951253354549408, -0.05628962442278862 ], "target_std": [ 2.026750326156616, 1.075918436050415 ] }, { "pred_mean": [ 0.0790666937828064, -0.0549687035381794 ], "pred_std": [ 1.9015724658966064, 0.9414774775505066 ], "target_mean": [ 0.028750987723469734, -0.007199820131063461 ], "target_std": [ 2.000044822692871, 1.0400867462158203 ] }, { "pred_mean": [ 0.1755228042602539, -0.08683786541223526 ], "pred_std": [ 2.0293917655944824, 1.0066801309585571 ], "target_mean": [ 0.13575679063796997, -0.049245692789554596 ], "target_std": [ 2.0310559272766113, 1.0684142112731934 ] }, { "pred_mean": [ 0.0953637957572937, 0.01983083039522171 ], "pred_std": [ 1.9573743343353271, 0.9901137948036194 ], "target_mean": [ 0.02048776112496853, 0.015579056926071644 ], "target_std": [ 2.0075297355651855, 1.0683743953704834 ] }, { "pred_mean": [ 0.009151271544396877, -0.024620816111564636 ], "pred_std": [ 1.9104688167572021, 0.9687676429748535 ], "target_mean": [ 0.07076805084943771, -0.020372634753584862 ], "target_std": [ 2.042053699493408, 1.0628199577331543 ] }, { "pred_mean": [ -0.15449641644954681, 0.07149290293455124 ], "pred_std": [ 1.8906241655349731, 0.9696427583694458 ], "target_mean": [ -0.13105426728725433, 0.08039151132106781 ], "target_std": [ 1.9721097946166992, 1.02797532081604 ] } ] }, "comparison": { "delta_mean": -1.2732128258794546, "idea_wins": 8, "n_pairs": 8, "per_seed_diffs": [ -1.3786678314208984, -1.113927498459816, -1.3399933725595474, -1.2648904770612717, -1.3451272547245026, -1.3624695241451263, -1.2300445139408112, -1.1505821347236633 ], "p_value": 0.0081, "mde": 0.08424709934543327, "mde_rel_pct": 5.862106372408011, "verdict": "idea better (significant)", "system_worked": true }, "mechanism_signature": { "track_structure": "endpoint Gaussian-mixture transport", "signature": { "prediction": "component smoothing should reduce target-label assignment variance", "predicted": 0.07482332631479949, "observed": 0.06172130233608186, "confirmed": false }, "custom_track": { "name": "latent_mixture_transport", "file": "custom_track.py", "domain": "diffusion-sampling" } } }