Vidya
completestandardruns3k/polar-dense: polar dense 4L/32H, 3000 updates (98 MB), 2.0490 bpb, 3.62 M MAC/byte, 2.14e+15 FLOPCORE 0.0089
2026-10-21 11:42 · max 500 examples/task · scorer v1 · bundle abc150c04fa3 · endpoint https://core.roland.blah.dev
| Task | Accuracy | Centered | Δ centered | n | Time |
|---|---|---|---|---|---|
| agi_eval_lsat_ar | 0.2783 | 0.0978 | — | 230 | 510s |
| arc_challenge | 0.2740 | 0.0320 | — | 500 | 947s |
| arc_easy | 0.3120 | 0.0827 | — | 500 | 858s |
| bigbench_cs_algorithms | 0.0000 | 0.0000 | — | 500 | 191s |
| bigbench_dyck_languages | 0.0340 | 0.0340 | — | 500 | 243s |
| bigbench_language_identification | 0.2360 | 0.1595 | — | 500 | 1146s |
| bigbench_operators | 0.0667 | 0.0667 | — | 210 | 79s |
| bigbench_qa_wikidata | 0.0000 | 0.0000 | — | 500 | 69s |
| bigbench_repeat_copy_logic | 0.0000 | 0.0000 | — | 32 | 15s |
| boolq | 0.3720 | -0.6526 | — | 500 | 567s |
| commonsense_qa | 0.4040 | 0.2550 | — | 500 | 856s |
| copa | 0.5600 | 0.1200 | — | 100 | 9s |
| coqa | 0.0000 | 0.0000 | — | 500 | 268s |
| hellaswag | 0.2340 | -0.0213 | — | 500 | 1068s |
| hellaswag_zeroshot | 0.2360 | -0.0187 | — | 500 | 202s |
| jeopardy | 0.0000 | 0.0000 | — | 500 | 208s |
| lambada_openai | 0.0000 | 0.0000 | — | 500 | 58s |
| openbook_qa | 0.2600 | 0.0133 | — | 500 | 66s |
| piqa | 0.4840 | -0.0320 | — | 500 | 425s |
| squad | 0.0000 | 0.0000 | — | 500 | 313s |
| winograd | 0.5201 | 0.0403 | — | 273 | 25s |
| winogrande | 0.5100 | 0.0200 | — | 500 | 49s |
Every example's outcome (per-option losses, greedy matches, correctness) is stored with this attempt.