# DRSR: Learning Set-Level Deletion Risk for Efficient Long-Horizon Agents

arXiv: https://arxiv.org/abs/2609.27276 · HTML: https://mlsanity.com/paper/2609.27276/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning.
- Builds on 42 of its 51 references that are in the map.
- First posted on arXiv 23 Sep 2026.

## Placed in

- LLM reasoning / [Reasoning for agents, planning and code](https://mlsanity.com/lr/reasoning-for-agents-planning-and-code/)
- LLM agents and tool use / [Agent memory systems](https://mlsanity.com/agents/agent-memory-systems/)

## Builds on (most cited first, 30 of 42)

- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Lost in the Middle: How Language Models Use Long Contexts](https://mlsanity.com/paper/2307.03172/) (2023)
- [SWE-bench: Can Language Models Resolve Real-World GitHub Issues?](https://mlsanity.com/paper/2310.06770/) (2023)
- [Efficient Streaming Language Models with Attention Sinks](https://mlsanity.com/paper/2309.17453/) (2023)
- [MemGPT: Towards LLMs as Operating Systems](https://mlsanity.com/paper/2310.08560/) (2023)
- [SnapKV: LLM Knows What You are Looking for Before Generation](https://mlsanity.com/paper/2404.14469/) (2024)
- [Large Language Models Can Be Easily Distracted by Irrelevant Context](https://mlsanity.com/paper/2302.00093/) (2023)
- [H2O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Models](https://mlsanity.com/paper/2306.14048/) (2023)
- [Semantic Uncertainty: Linguistic Invariances for Uncertainty Estimation in Natural Language Generation](https://mlsanity.com/paper/2302.09664/) (2023)
- [AppWorld: A Controllable World of Apps and People for Benchmarking Interactive Coding Agents](https://mlsanity.com/paper/2407.18901/) (2024)
- [SelectiveNet: A Deep Neural Network with an Integrated Reject Option](https://mlsanity.com/paper/1901.09192/) (2019)
- [LLMLingua-2: Data Distillation for Efficient and Faithful Task-Agnostic Prompt Compression](https://mlsanity.com/paper/2403.12968/) (2024)
- [Conformal Risk Control](https://mlsanity.com/paper/2208.02814/) (2022)
- [Chain of Draft: Thinking Faster by Writing Less](https://mlsanity.com/paper/2502.18600/) (2025)
- [Distribution-Free, Risk-Controlling Prediction Sets](https://mlsanity.com/paper/2101.02703/) (2021)
- [LLMLingua: Compressing Prompts for Accelerated Inference of Large Language Models](https://mlsanity.com/paper/2310.05736/) (2023)
- [Consistent Individualized Feature Attribution for Tree Ensembles](https://mlsanity.com/paper/1802.03888/) (2018)
- [TokenSkip: Controllable Chain-of-Thought Compression in LLMs](https://mlsanity.com/paper/2502.12067/) (2025)
- [Ada-KV: Optimizing KV Cache Eviction by Adaptive Budget Allocation for Efficient LLM Inference](https://mlsanity.com/paper/2407.11550/) (2024)
- [Unifying distillation and privileged information](https://mlsanity.com/paper/1511.03643/) (2015)
- [Explaining by Removing: A Unified Framework for Model Explanation](https://mlsanity.com/paper/2011.14878/) (2020)
- [Kernel Language Entropy: Fine-grained Uncertainty Quantification for LLMs from Semantic Similarities](https://mlsanity.com/paper/2405.20003/) (2024)
- [Scaling Long-Horizon LLM Agent via Context-Folding](https://mlsanity.com/paper/2510.11967/) (2025)
- [The Shapley Taylor Interaction Index](https://mlsanity.com/paper/1902.05622/) (2019)
- [ACON: Optimizing Context Compression for Long-horizon LLM Agents](https://mlsanity.com/paper/2510.00615/) (2025)
- [Language Models with Conformal Factuality Guarantees](https://mlsanity.com/paper/2402.10978/) (2024)
- [AgentFold: Long-Horizon Web Agents with Proactive Context Management](https://mlsanity.com/paper/2510.24699/) (2025)
- [Faith-Shap: The Faithful Shapley Interaction Index](https://mlsanity.com/paper/2203.00870/) (2022)
- [Memory as Action: Autonomous Context Curation for Long-Horizon Agentic Tasks](https://mlsanity.com/paper/2510.12635/) (2025)

## Abstract (arXiv)

Long-horizon language-model agents accumulate reasoning traces, tool exchanges, and observations whose relevance changes with the current decision. Existing compression strategies often score historical units independently, but the safety of deleting several units is generally not determined by their singleton scores: redundant evidence, accumulated small effects, and the information that remains after deletion all matter. We introduce Direct Relational Set-Risk Pruning (DRSR), which formulates agent-history compression as risk-constrained selection over deletion sets. Offline, DRSR constructs exact counterfactual supervision by jointly deleting protocol-valid history Blocks and measuring the change in teacher-forced likelihood of the same recorded next output. A lightweight scorer then predicts set-level harm from online-visible relations between candidate history and the current pre-action state, together with deleted-retained and pairwise set structure. At deployment, DRSR evaluates a small set of structurally valid deletion candidates with the lightweight scorer and removes the largest feasible set under recency, protocol, budget, and learned-risk constraints, abstaining when no set is sufficiently safe. On WorkBuddyBench Full260, DRSR increases mean reward from 0.699 to 0.802 while reducing total model tokens by 20.820%. On the fixed Eval40 comparison, it obtains 0.794 reward at 1.211M tokens per task, using 35.850% fewer tokens than the uncompressed agent. Mechanistic analyses and ablations further show that decision-conditioned relations, retained-context information, pair interactions, and abstention each contribute to reliable pruning.