Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | https://arxiv.org/abs/2603.23420 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866912980817412096 |
|---|---|
| author | Qu, Yaonan Lu, Meng |
| author_facet | Qu, Yaonan Lu, Meng |
| contents | If autoresearch is itself a form of research, then autoresearch can be applied to research itself. We take this idea literally: we use an autoresearch loop to optimize the autoresearch loop. Every existing autoresearch system -- from Karpathy's single-track loop to AutoResearchClaw's multi-batch extension and EvoScientist's persistent memory -- was improved by a human who read the code, identified a bottleneck, and wrote new code. We ask whether an LLM can do the same, autonomously. We present Bilevel Autoresearch, a bilevel framework where an outer loop meta-optimizes the inner autoresearch loop by generating and injecting new search mechanisms as Python code at runtime. The inner loop optimizes the task; the outer loop optimizes how the inner loop searches. Both loops use the same LLM -- no stronger model is needed at the meta level. On Karpathy's GPT pretraining benchmark, the meta-autoresearch outer loop achieves a 5x improvement over the standard inner loop alone (-0.045 vs. -0.009 val_bpb), while parameter-level adjustment without mechanism change yields no reliable gain. The outer loop autonomously discovers mechanisms from combinatorial optimization, multi-armed bandits, and design of experiments -- without human specification of which domains to explore. These mechanisms succeed by breaking the inner loop's deterministic search patterns, forcing exploration of directions the LLM's priors systematically avoid. The core principle is simple: if autoresearch can meta-autoresearch itself, it can, in principle, meta-autoresearch anything with a measurable objective. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_23420 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Bilevel Autoresearch: Meta-Autoresearching Itself Qu, Yaonan Lu, Meng Artificial Intelligence If autoresearch is itself a form of research, then autoresearch can be applied to research itself. We take this idea literally: we use an autoresearch loop to optimize the autoresearch loop. Every existing autoresearch system -- from Karpathy's single-track loop to AutoResearchClaw's multi-batch extension and EvoScientist's persistent memory -- was improved by a human who read the code, identified a bottleneck, and wrote new code. We ask whether an LLM can do the same, autonomously. We present Bilevel Autoresearch, a bilevel framework where an outer loop meta-optimizes the inner autoresearch loop by generating and injecting new search mechanisms as Python code at runtime. The inner loop optimizes the task; the outer loop optimizes how the inner loop searches. Both loops use the same LLM -- no stronger model is needed at the meta level. On Karpathy's GPT pretraining benchmark, the meta-autoresearch outer loop achieves a 5x improvement over the standard inner loop alone (-0.045 vs. -0.009 val_bpb), while parameter-level adjustment without mechanism change yields no reliable gain. The outer loop autonomously discovers mechanisms from combinatorial optimization, multi-armed bandits, and design of experiments -- without human specification of which domains to explore. These mechanisms succeed by breaking the inner loop's deterministic search patterns, forcing exploration of directions the LLM's priors systematically avoid. The core principle is simple: if autoresearch can meta-autoresearch itself, it can, in principle, meta-autoresearch anything with a measurable objective. |
| title | Bilevel Autoresearch: Meta-Autoresearching Itself |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2603.23420 |