A LoRA-fine-tuned Qwen 3.5 2B model for task-conditioned tool-output pruning reaches 0.86 recall and 0.80 F1 on a new 618-example test set while removing 92% of input tokens and outperforming larger zero-shot models.
Tim Dettmers, Artidoro Pagnoni, Ari Holtzman, and Luke Zettlemoyer
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
abstract
Retrieval-augmented generation improves various aspects of large language models (LLMs) generation, but suffers from computational overhead caused by long contexts as well as the propagation of irrelevant retrieved information into generated responses. Context pruning deals with both aspects, by removing irrelevant parts of retrieved contexts before LLM generation. Existing context pruning approaches are however limited, and do not provide a universal model that would be both efficient and robust in a wide range of scenarios, e.g., when contexts contain a variable amount of relevant information or vary in length, or when evaluated on various domains. In this work, we close this gap and introduce Provence (Pruning and Reranking Of retrieVEd relevaNt ContExts), an efficient and robust context pruner for Question Answering, which dynamically detects the needed amount of pruning for a given context and can be used out-of-the-box for various domains. The three key ingredients of Provence are formulating the context pruning task as sequence labeling, unifying context pruning capabilities with context reranking, and training on diverse data. Our experimental results show that Provence enables context pruning with negligible to no drop in performance, in various domains and settings, at almost no cost in a standard RAG pipeline. We also conduct a deeper analysis alongside various ablations to provide insights into training context pruners for future work.
years
2026 5representative citing papers
KV-cache eviction, prompt compression, recurrent state bounding, and agent memory consolidation are unified as one rate-distortion problem with a shared lower bound, shared failure mode, and transferable mechanisms.
GroundedCache reduces unsafe-served rate in RAG answer caching to 0-1.5% (vs 15-51.5% naive) via four validation gates while keeping p50 latency within 1.07x of no-cache baseline.
EASE-TTT creates a soft attention target from evidence chunks to guide query-side test-time adaptation, yielding higher macro-average scores than full-context, retrieval-only, and standard qTTT baselines on six LongBench QA tasks.
LongAttnComp adapts attention-based context compression with token-level chunking, positional reordering, and two-stage fine-tuning to match full-context performance on InfiniteBench Code-Debug and improve multi-document reasoning on LongBench v2 across model families.
citing papers explorer
-
Squeez: Task-Conditioned Tool-Output Pruning for Coding Agents
A LoRA-fine-tuned Qwen 3.5 2B model for task-conditioned tool-output pruning reaches 0.86 recall and 0.80 F1 on a new 618-example test set while removing 92% of input tokens and outperforming larger zero-shot models.
-
What to Keep, What to Forget: A Rate--Distortion View of Memory Compaction in LLMs and Agents
KV-cache eviction, prompt compression, recurrent state bounding, and agent memory consolidation are unified as one rate-distortion problem with a shared lower bound, shared failure mode, and transferable mechanisms.
-
Grounded Cache Routing for Retrieval-Augmented Generation: When Is It Safe to Reuse an Answer?
GroundedCache reduces unsafe-served rate in RAG answer caching to 0-1.5% (vs 15-51.5% naive) via four validation gates while keeping p50 latency within 1.07x of no-cache baseline.
-
EASE-TTT: Evidence-Aligned Selective Test-Time Training for Long-Context Question Answering
EASE-TTT creates a soft attention target from evidence chunks to guide query-side test-time adaptation, yielding higher macro-average scores than full-context, retrieval-only, and standard qTTT baselines on six LongBench QA tasks.
-
LongAttnComp: Cross-Family Context Compression for Long-Context Reasoning
LongAttnComp adapts attention-based context compression with token-level chunking, positional reordering, and two-stage fine-tuning to match full-context performance on InfiniteBench Code-Debug and improve multi-document reasoning on LongBench v2 across model families.