Proved most LLMs were undertrained, reshaped scaling strategy industry-wide
| Benchmark | Score | Rank |
|---|---|---|
HellaSwag Common sense reasoning about everyday situations | 82.3% | #35 / 36 |
ARC-C Grade-school science questions requiring reasoning | 83.7% | #39 / 40 |
HumanEval Coding ability - generating correct Python functions | 31.8% | #47 / 49 |
MMLU Tests knowledge across 57 subjects from STEM to humanities | 67.5% | #52 / 53 |