Research map / Large language models / Retrieval and knowledge
Demystifying Deep Search: A Holistic Evaluation with Hint-free Multi-Hop Questions and Factorised Metrics
Maojia Song, Liu Renhang, Xinyu Wang, Yong Jiang, Pengjun Xie, Fei Huang, Soujanya Poria, Jingren Zhou
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Wrap DeepSeek-R1 in EvidenceLoop (evidence memory, verification, parallel rounds) | DeepSeek-R1 · WebDetective | Pass@1: 20.0 -> 25.0 | Improved |
| Wrap DeepSeek-R1 in EvidenceLoop | DeepSeek-R1 · WebDetective | Knowledge Utilization F1: 15.55 -> 23.79 | Improved |
| Wrap DeepSeek-R1 in EvidenceLoop | DeepSeek-R1 · WebDetective | Good Refusal F1: 18.81 -> 17.98 | No clear effect |
| Wrap DeepSeek-R1 in EvidenceLoop | DeepSeek-R1 · WebDetective | Search Score: 65.50 -> 62.50 | Hurt |
| Enlarge context window from 8K to 32K (ablation) | Claude-Opus-4.1 · WebDetective | Pass@1: plateaus at about 50% | No clear effect |
| Increase EvidenceLoop breadth from 1 to 3 (iteration=2) (ablation) | EvidenceLoop · WebDetective | Pass@1: 49 -> 56 | Improved |
| Increase EvidenceLoop breadth from 1 to 3 (iteration=2) (ablation) | EvidenceLoop · WebDetective | Generation Score: 21 -> 21 | No clear effect |
| Increase EvidenceLoop breadth from 1 to 3 (iteration=2) (ablation) | EvidenceLoop · WebDetective | Search Score: 45 -> 46 | No clear effect |
| Higher search score (retrieval) on frontier models | GPT-5 vs Grok-4 · WebDetective | Generation Score: 80.00 vs 77.50 search; 23.21 vs 34.71 generation | Mixed |
Closest papers by meaning
- From Intent to Evidence: A Categorical Approach for Structural Evaluation of Deep Research Agents (NeurIPS 2026 · similarity 0.90)
- AgentHop: A Diagnostic Benchmark for Agentic Multi-Hop Scientific Question Answering (NeurIPS 2026 · similarity 0.90)
- MC-Search: Evaluating and Enhancing Multimodal Agentic Search with Structured Long Reasoning Chains (ICLR 2026 · similarity 0.89)
- Empowering Efficiency and Efficacy in WebAgent via Enabling Info-Rich Seeking (ICLR 2026 · similarity 0.89)
- Mind2Web 2: Evaluating Agentic Search with Agent-as-a-Judge (NeurIPS 2025 · similarity 0.89)
- Fathom-DeepResearch: Unlocking Long Horizon Information Retrieval and Synthesis for SLMs (ICLR 2026 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
