Agentic Search in the Wild: Intents and Trajectory Dynamics from 14M+ Real Search Requests

Forum topic · 小凯 · 2026-07-05

Summary

This arXiv paper (2601.17617) presents a large-scale log analysis of LLM-powered agentic search based on 14.44M search requests across 3.97M sessions collected from DeepResearchGym, an open-source search API accessed by external agentic clients. The authors sessionize logs, assign session-level intents and step-wise query-reformulation labels via LLM-based annotation, and propose Context-driven Term Adoption Rate (CTAR) to measure whether newly introduced query terms are lexically traceable to previously retrieved evidence. Key findings: over 90% of multi-turn sessions contain at most ten steps and 89% of inter-step intervals are under one minute; fact-seeking sessions show increasing repetition while reasoning-heavy sessions sustain broader exploration; on average 54% of new query terms appear in the accumulated evidence context, with traceability extending to earlier retrieval steps. These results suggest candidate signals for repetition-aware stopping, intent-adaptive retrieval budgeting, and cross-step context tracking. Anonymized logs are released on HuggingFace.

Overview

Agentic Search in the Wild: Intents and Trajectory Dynamics from 14M+ Real Search Requests is an academic paper (arXiv: 2601.17617, dated 2026-01-24) by Jingjie Ning, João Coelho, Yibo Kong, Yunfan Long, Bruno Martins, João Magalhães, et al. (8 authors total).

Background

LLM-powered search agents are increasingly used for multi-step information seeking, yet the IR community lacks empirical understanding of how agentic search sessions actually unfold and how retrieved evidence shapes later queries. This paper addresses that gap with real-world logs.

Methodology

Data: 14.44M search requests (3.97M sessions) from DeepResearchGym, an open-source search API used by external agentic clients.
Annotation: Logs are sessionized; session-level intents and step-wise query-reformulation labels are assigned using LLM-based annotation.
Metric: The paper proposes Context-driven Term Adoption Rate (CTAR) to quantify whether newly introduced query terms are lexically traceable to previously retrieved evidence.

Key Findings

1. Session shape: Over 90% of multi-turn sessions contain at most ten steps; 89% of inter-step intervals fall under one minute. 2. Intent-dependent behavior: Fact-seeking sessions exhibit high repetition that increases over time, while sessions requiring reasoning sustain broader exploration. 3. Evidence-driven reformulation: On average, 54% of newly introduced query terms appear in the accumulated evidence context, with additional traceability to earlier steps beyond the most recent retrieval.

Implications

The findings provide candidate signals for:

Repetition-aware stopping criteria
Intent-adaptive retrieval budgeting
Explicit cross-step context tracking

Resources

Paper: arXiv:2601.17617
Anonymized logs released publicly: HuggingFace dataset

Editorial Notes

This entry fits within the Agentic Search track, cross-referencing related work such as *Agentic Information Retrieval* (arXiv:2410.09713), *AceSearcher* (arXiv:2509.24193), and surveys on synergizing RAG with reasoning (arXiv:2504.15909). For engineering teams, the behavioral statistics (session depth, inter-step latency, term adoption) offer practical defaults for budgeting agentic retrieval pipelines.