Guardado en:
| Autores principales: | , , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2505.13652 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866908371231178752 |
|---|---|
| author | Zainullina, Karina Golubev, Alexander Trofimova, Maria Polezhaev, Sergei Badertdinov, Ibragim Litvintseva, Daria Karasik, Simon Fisin, Filipp Skvortsov, Sergei Nekrashevich, Maksim Shevtsov, Anton Yangel, Boris |
| author_facet | Zainullina, Karina Golubev, Alexander Trofimova, Maria Polezhaev, Sergei Badertdinov, Ibragim Litvintseva, Daria Karasik, Simon Fisin, Filipp Skvortsov, Sergei Nekrashevich, Maksim Shevtsov, Anton Yangel, Boris |
| contents | Large language models (LLMs) have recently achieved remarkable results in complex multi-step tasks, such as mathematical reasoning and agentic software engineering. However, they often struggle to maintain consistent performance across multiple solution attempts. One effective approach to narrow the gap between average-case and best-case performance is guided test-time search, which explores multiple solution paths to identify the most promising one. Unfortunately, effective search techniques (e.g. MCTS) are often unsuitable for non-serializable RL environments, such as Docker containers, where intermediate environment states cannot be easily saved and restored. We investigate two complementary search strategies applicable to such environments: 1-step lookahead and trajectory selection, both guided by a learned action-value function estimator. On the SWE-bench Verified benchmark, a key testbed for agentic software engineering, we find these methods to double the average success rate of a fine-tuned Qwen-72B model, achieving 40.8%, the new state-of-the-art for open-weights models. Additionally, we show that these techniques are transferable to more advanced closed models, yielding similar improvements with GPT-4o. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_13652 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Guided Search Strategies in Non-Serializable Environments with Applications to Software Engineering Agents Zainullina, Karina Golubev, Alexander Trofimova, Maria Polezhaev, Sergei Badertdinov, Ibragim Litvintseva, Daria Karasik, Simon Fisin, Filipp Skvortsov, Sergei Nekrashevich, Maksim Shevtsov, Anton Yangel, Boris Software Engineering Computation and Language Large language models (LLMs) have recently achieved remarkable results in complex multi-step tasks, such as mathematical reasoning and agentic software engineering. However, they often struggle to maintain consistent performance across multiple solution attempts. One effective approach to narrow the gap between average-case and best-case performance is guided test-time search, which explores multiple solution paths to identify the most promising one. Unfortunately, effective search techniques (e.g. MCTS) are often unsuitable for non-serializable RL environments, such as Docker containers, where intermediate environment states cannot be easily saved and restored. We investigate two complementary search strategies applicable to such environments: 1-step lookahead and trajectory selection, both guided by a learned action-value function estimator. On the SWE-bench Verified benchmark, a key testbed for agentic software engineering, we find these methods to double the average success rate of a fine-tuned Qwen-72B model, achieving 40.8%, the new state-of-the-art for open-weights models. Additionally, we show that these techniques are transferable to more advanced closed models, yielding similar improvements with GPT-4o. |
| title | Guided Search Strategies in Non-Serializable Environments with Applications to Software Engineering Agents |
| topic | Software Engineering Computation and Language |
| url | https://arxiv.org/abs/2505.13652 |