Language Modeling on FineWeb Nanochat 5M token (val)
2.681BPB (val)AutoResearch-RL
Evaluation Results
| Method | Links | |
|---|---|---|
| AutoResearch-RL# Expts=1012026.03 | 2.681 | |
| Greedy LLM# Expts=88, RL fine-tuning=none, mode=zero-shot2026.03 | 2.734 | |
| Random Search# Expts=932026.03 | 2.791 | |
| Human Expert# Expts=1, model=GPT-2-small (124M)2026.03 | 2.847 |