Vidya
completesmokeruns/cpu-polar: polar dense 3L/16H, 2000 updates, 2.3583 bpb, 0.78 M MAC/byte, 1.54e+14 FLOPCORE -0.0138
2026-10-21 08:52 · max 50 examples/task · scorer v1 · bundle abc150c04fa3 · endpoint https://core.roland.blah.dev
| Task | Accuracy | Centered | Δ centered | n | Time |
|---|---|---|---|---|---|
| agi_eval_lsat_ar | 0.2200 | 0.0250 | — | 50 | 59s |
| arc_challenge | 0.2800 | 0.0400 | — | 50 | 53s |
| arc_easy | 0.3200 | 0.0933 | — | 50 | 46s |
| bigbench_cs_algorithms | 0.0000 | 0.0000 | — | 50 | 13s |
| bigbench_dyck_languages | 0.0000 | 0.0000 | — | 50 | 12s |
| bigbench_language_identification | 0.2200 | 0.1419 | — | 50 | 59s |
| bigbench_operators | 0.0200 | 0.0200 | — | 50 | 13s |
| bigbench_qa_wikidata | 0.0000 | 0.0000 | — | 50 | 6s |
| bigbench_repeat_copy_logic | 0.0000 | 0.0000 | — | 32 | 9s |
| boolq | 0.3800 | -0.6316 | — | 50 | 35s |
| commonsense_qa | 0.1200 | -0.1000 | — | 50 | 49s |
| copa | 0.5800 | 0.1600 | — | 50 | 4s |
| coqa | 0.0000 | 0.0000 | — | 50 | 15s |
| hellaswag | 0.2600 | 0.0133 | — | 50 | 59s |
| hellaswag_zeroshot | 0.2400 | -0.0133 | — | 50 | 13s |
| jeopardy | 0.0000 | 0.0000 | — | 50 | 12s |
| lambada_openai | 0.0000 | 0.0000 | — | 50 | 4s |
| openbook_qa | 0.2400 | -0.0133 | — | 50 | 4s |
| piqa | 0.5200 | 0.0400 | — | 50 | 29s |
| squad | 0.0000 | 0.0000 | — | 50 | 20s |
| winograd | 0.5200 | 0.0400 | — | 50 | 4s |
| winogrande | 0.4400 | -0.1200 | — | 50 | 4s |
Every example's outcome (per-option losses, greedy matches, correctness) is stored with this attempt.