aiqbench analysis
GrepSeek advances LLM search agents by replacing traditional ranked retrieval with direct corpus interaction via executable shell commands, significantly reducing latency while maintaining accuracy. The system stabilizes reinforcement learning through a two-stage training process involving an answer-aware Tutor and Group Relative Policy Optimization to handle large-scale text corpora effectively. This approach offers a novel paradigm for knowledge-intensive tasks where the agent actively navigates the file system rather than relying on pre-built indexes.
GrepSeek: Training Search Agents for Direct Corpus Interaction
arXiv:2605.29307v2 Announce Type: replace-cross Abstract: Large Language Model (LLM) search agents have shown strong promise on knowledge-intensive tasks through iterative reasoning and retrieval. Most existing systems rely on retrievers that return ranked documents from a pre-built index. We explore a complementary paradigm in which the agent treats the corpus as the search environment and finds evidence through executable shell commands. We introduce GrepSeek, an optimized direct corpus interaction (DCI) agent that learns to find, filter, and compose evidence over large text corpora. To stabilize reinforcement learning (RL) over large corpora, we train in two stages: first, we initialize the policy using verified, causally grounded search trajectories generated by an answer-aware Tutor and an answer-blind Planner; then, we refine the policy using Group Relative Policy Optimization (GRPO). To make DCI practical at scale, we introduce two semantics-preserving execution optimizations: Pruned Adaptive Command Execution, which reduces shell-based search latency by up to $77\times$ on a 14GB corpus with 21 million documents using a compact auxiliary structure, and Sharded-Parallel Corpus Search, which achieves up to $7.6\times$ speedup without additional preprocessing; both preserve equivalence with sequential execution. Across eight open-domain QA benchmarks, GrepSeek achieves the strongest overall performance, with a statistically significant relative improvement of $5.7\%$ over the best baseline. Our analysis shows how DCI-optimized agents conduct flexible and effective compositional search through direct corpus interaction.
Abstract:Large Language Model (LLM) search agents have shown strong promise on knowledge-intensive tasks through iterative reasoning and retrieval. Most existing systems rely on retrievers that return ranked documents from a pre-built index. We explore a complementary paradigm in which the agent treats the corpus as the search environment and finds evidence through executable shell commands. We introduce GrepSeek, an optimized direct corpus interaction (DCI) agent that learns to find, filter, and compose evidence over large text corpora. To stabilize reinforcement learning (RL) over large corpora, we train in two stages: first, we initialize the policy using verified, causally grounded search trajectories generated by an answer-aware Tutor and an answer-blind Planner; then, we refine the policy using Group Relative Policy Optimization (GRPO). To make DCI practical at scale, we introduce two semantics-preserving execution optimizations: Pruned Adaptive Command Execution, which reduces shell-based search latency by up to $77\times$ on a 14GB corpus with 21 million documents using a compact auxiliary structure, and Sharded-Parallel Corpus Search, which achieves up to $7.6\times$ speedup without additional preprocessing; both preserve equivalence with sequential execution. Across eight open-domain QA benchmarks, GrepSeek achieves the strongest overall performance, with a statistically significant relative improvement of $5.7\%$ over the best baseline. Our analysis shows how DCI-optimized agents conduct flexible and effective compositional search through direct corpus interaction.
Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR); Machine Learning (cs.LG)
arXiv:2605.29307 [cs.CL]
(or arXiv:2605.29307v2 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2605.29307 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 319 words
Related coverage
Dr-DCI: Scaling Direct Corpus Interaction via Dynamic Workspace Expansion
arXiv · Jun 17, 2026
The Continual Constitutional Convention: How language models revise the documents that govern them
AI Alignment Forum · Oct 13, 2026
Research Note: Filtering Subversion-Relevant Information From Pretraining Data Is Feasible
LessWrong · Oct 5, 2026
Presentation: Automating the Web With MCP: Infra That Doesn’t Break
InfoQ · Jun 16, 2026