Towards Retrieving Interaction Spaces for Agentic Search</p>\n","updatedAt":"2026-06-08T13:24:02.843Z","author":{"_id":"60d97add9fe99457e2010efe","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1624866593786-60d97add9fe99457e2010efe.png","fullname":"Shengyao Zhuang","name":"ArvinZhuang","type":"user","isPro":true,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":7,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.5320842266082764},"editors":["ArvinZhuang"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1624866593786-60d97add9fe99457e2010efe.png"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2606.06880","authors":[{"_id":"6a26c241da05d61ad5d10d02","name":"Shengyao Zhuang","hidden":false},{"_id":"6a26c241da05d61ad5d10d03","name":"Yuansheng Ni","hidden":false},{"_id":"6a26c241da05d61ad5d10d04","name":"Hengxin Fun","hidden":false},{"_id":"6a26c241da05d61ad5d10d05","name":"Jimmy Lin","hidden":false},{"_id":"6a26c241da05d61ad5d10d06","name":"Xueguang Ma","hidden":false}],"publishedAt":"2026-06-05T00:00:00.000Z","submittedOnDailyAt":"2026-06-08T00:00:00.000Z","title":"Towards Retrieving Interaction Spaces for Agentic Search","submittedOnDailyBy":{"_id":"60d97add9fe99457e2010efe","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1624866593786-60d97add9fe99457e2010efe.png","isPro":true,"fullname":"Shengyao Zhuang","user":"ArvinZhuang","type":"user","name":"ArvinZhuang"},"summary":"Retrieval for search agents is still inherited from non-agentic information retrieval: a retriever ranks the corpus and the agent reads a small set of returned documents. Recent direct corpus interaction (DCI) work shows that agents can instead interact with the raw corpus through shell tools such as grep and file reads. But unbounded interaction does not scale: every broad shell command is a scan over the whole corpus, and latency degrades sharply as the corpus grows. We argue that the role of retrieval for agentic search is not just to select documents that fit in the LLM context window, but to construct an interaction space: a bounded subset of the corpus the agent can explore with associated tools. Two design consequences follow. The space needs a boundary supplied by retrieval, and the objects within it should be processed for interaction. As a proof of concept, we propose RISE (Retrieving Interaction SpacE): we use BM25 to construct the interaction space; meanwhile, its documents are processed during indexing for shell-style navigation. On BrowseComp-Plus, RISE matches the pure-shell DCI baseline at 78% accuracy with gpt-5.4-mini at roughly one quarter of the per-query cost. At 1M documents, RISE-BM25 reaches 81% on gpt-5.4-mini, whereas DCI on gpt-5.4-nano degrades to 60% with 33 of 100 wall-clock failures.","upvotes":1,"discussionId":"6a26c241da05d61ad5d10d07","githubRepo":"https://github.com/texttron/RISE","githubRepoAddedBy":"user","ai_summary":"RISE framework constructs bounded interaction spaces for agentic search by combining BM25 retrieval with preprocessed document indexing to enable efficient corpus exploration while maintaining high accuracy at scale.","ai_keywords":["retrieval","direct corpus interaction","BM25","interaction space","corpus scaling","shell tools","document processing","RISE","BrowseComp-Plus"],"ai_summary_model":"Qwen/Qwen2.5-Coder-32B-Instruct","githubStars":2},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"60d97add9fe99457e2010efe","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1624866593786-60d97add9fe99457e2010efe.png","isPro":true,"fullname":"Shengyao Zhuang","user":"ArvinZhuang","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2606/2606.06880.md"}">
Towards Retrieving Interaction Spaces for Agentic Search
Abstract
RISE framework constructs bounded interaction spaces for agentic search by combining BM25 retrieval with preprocessed document indexing to enable efficient corpus exploration while maintaining high accuracy at scale.
Retrieval for search agents is still inherited from non-agentic information retrieval: a retriever ranks the corpus and the agent reads a small set of returned documents. Recent direct corpus interaction (DCI) work shows that agents can instead interact with the raw corpus through shell tools such as grep and file reads. But unbounded interaction does not scale: every broad shell command is a scan over the whole corpus, and latency degrades sharply as the corpus grows. We argue that the role of retrieval for agentic search is not just to select documents that fit in the LLM context window, but to construct an interaction space: a bounded subset of the corpus the agent can explore with associated tools. Two design consequences follow. The space needs a boundary supplied by retrieval, and the objects within it should be processed for interaction. As a proof of concept, we propose RISE (Retrieving Interaction SpacE): we use BM25 to construct the interaction space; meanwhile, its documents are processed during indexing for shell-style navigation. On BrowseComp-Plus, RISE matches the pure-shell DCI baseline at 78% accuracy with gpt-5.4-mini at roughly one quarter of the per-query cost. At 1M documents, RISE-BM25 reaches 81% on gpt-5.4-mini, whereas DCI on gpt-5.4-nano degrades to 60% with 33 of 100 wall-clock failures.
Community
Towards Retrieving Interaction Spaces for Agentic Search
Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images
Cite arxiv.org/abs/2606.06880 in a model README.md to link it from this page.
Cite arxiv.org/abs/2606.06880 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2606.06880 in a Space README.md to link it from this page.
Discussion (0)
Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.
Sign in →No comments yet. Sign in and be the first to say something.