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.
Title resolution pending
8 Pith papers cite this work. Polarity classification is still indexing.
8
Pith papers citing it
citation-role summary
background 2
citation-polarity summary
years
2026 8roles
background 2polarities
background 2representative citing papers
DLH and DLH-M use distance-aware labeling sets compressed via hashing to support graph range filters in filtered ANN, delivering up to 70.3% higher throughput at over 98.5% recall.
MAGE uses an agentic shadow memory to proactively detect and mitigate long-horizon threats in LLM agents by distilling safety context and assessing action risks before execution.
HingeMem segments dialogue memory via boundary-triggered hyperedges over four elements and applies query-adaptive retrieval, yielding ~20% relative gains and 68% lower QA token cost versus baselines on LOCOMO.
CiteAudit supplies a human-validated benchmark and multi-agent verification system that outperforms existing LLMs and commercial tools at detecting hallucinated scientific references.
A framework distills noisy user logs into rules and preference pairs, clusters them by query and feedback, then fine-tunes either an expert adapter or a critic adapter to improve future responses.
MemCoT transforms long-context LLM reasoning into an iterative stateful search using multi-view memory for evidence localization and dual short-term memory for guiding decisions, achieving SOTA on LoCoMo and LongMemEval-S benchmarks.
citing papers explorer
-
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.
-
Approximate Nearest Neighbor Search with Graph Range Filters
DLH and DLH-M use distance-aware labeling sets compressed via hashing to support graph range filters in filtered ANN, delivering up to 70.3% higher throughput at over 98.5% recall.
-
MAGE: Safeguarding LLM Agents against Long-Horizon Threats via Shadow Memory
MAGE uses an agentic shadow memory to proactively detect and mitigate long-horizon threats in LLM agents by distilling safety context and assessing action risks before execution.
-
HingeMem: Boundary Guided Long-Term Memory with Query Adaptive Retrieval for Scalable Dialogues
HingeMem segments dialogue memory via boundary-triggered hyperedges over four elements and applies query-adaptive retrieval, yielding ~20% relative gains and 68% lower QA token cost versus baselines on LOCOMO.
-
CiteAudit: You Cited It, But Did You Read It? A Benchmark for Verifying Scientific References in the LLM Era
CiteAudit supplies a human-validated benchmark and multi-agent verification system that outperforms existing LLMs and commercial tools at detecting hallucinated scientific references.
-
Improve Large Language Model Systems with User Logs
A framework distills noisy user logs into rules and preference pairs, clusters them by query and feedback, then fine-tunes either an expert adapter or a critic adapter to improve future responses.
-
MemCoT: Test-Time Scaling through Memory-Driven Chain-of-Thought
MemCoT transforms long-context LLM reasoning into an iterative stateful search using multi-view memory for evidence localization and dual short-term memory for guiding decisions, achieving SOTA on LoCoMo and LongMemEval-S benchmarks.
- Memory in the LLM Era: Modular Architectures and Strategies in a Unified Framework