{"arxiv_id":"2603.18718","incoming":{"papers":6,"sentences":6,"groups":[{"kind":"background","label":"Cites it as background","papers":2,"sentences":2,"quotes":[{"arxiv_id":"2608.00017","title":"Memory Reward Inflation in Self-Improving LLM Agents","url":"/paper/2608.00017","section":"Related Work","sentence":"Adversarial memory-poisoning work studies settings where external attackers inject misleading information into agent memory [ref] .","intent_p":0.99,"contests_p":null},{"arxiv_id":"2608.20991","title":"Trojaning the Alignment: Stealthy Backdoor Attacks against Graph Foundation Models","url":"/paper/2608.20991","section":"Introduction","sentence":"Graph Foundation Models (GFMs) are increasingly used to learn graph representations that transfer across tasks, domains, and data regimes [ref] .","intent_p":0.93,"contests_p":null}]}],"unlabelled_sentences":4},"outgoing":{"papers":14,"sentences":32,"groups":[{"kind":"method_or_tool","label":"Uses the method or tool of","papers":4,"sentences":5,"quotes":[{"arxiv_id":"2508.19828","title":"Memory-R1: Enhancing Large Language Model Agents to Manage and Utilize Memories via Reinforcement Learning","url":"/paper/2508.19828","section":"Flexibility across Storage Backends","sentence":"To assess the flexibility of across storage backends, we instantiate it on top of three memory systems: Single-Agent [ref] ($_SA$), A-Mem ($_AM$), and LightMem ($_LM$).","intent_p":0.92,"contests_p":null},{"arxiv_id":"2410.21276","title":"GPT-4o System Card","url":"/paper/2410.21276","section":"Motivating Analysis: Strategic Blindness","sentence":"We compare three progressively stronger baselines using GPT-4o-mini [ref] as the backbone: (i) Static, which performs memory construction followed by one-shot top-$30$ retrieval; (ii) Unguided Active, which adds iterative query rewriting without strategic guidance; and (iii) Strategic Active, which introduces a planner to guide both construction and retrieval.","intent_p":0.87,"contests_p":null},{"arxiv_id":"2502.12110","title":"A-MEM: Agentic Memory for LLM Agents","url":"/paper/2502.12110","section":"Reasoning-Aware Coordination over the Forward Path","sentence":"Given $c_t$, $M̃_t-1$, and $g_t^S$, it selects an action $a_t^S ∈{, , , }$: [equation] The guidance signal $g_t^S$ helps $π_s$ filter noise, consolidate redundancy, and resolve conflicts at the source rather than blindly appending. $π_s$ is backend-agnostic and can wrap diverse memory implementations such as LightMem [ref] and A-Mem [ref] .","intent_p":0.86,"contests_p":null},{"arxiv_id":"2510.18866","title":"LightMem: Lightweight and Efficient Memory-Augmented Generation","url":"/paper/2510.18866","section":"Reasoning-Aware Coordination over the Forward Path","sentence":"Given $c_t$, $M̃_t-1$, and $g_t^S$, it selects an action $a_t^S ∈{, , , }$: [equation] The guidance signal $g_t^S$ helps $π_s$ filter noise, consolidate redundancy, and resolve conflicts at the source rather than blindly appending. $π_s$ is backend-agnostic and can wrap diverse memory implementations such as LightMem [ref] and A-Mem [ref] .","intent_p":0.86,"contests_p":null}]},{"kind":"data_used","label":"Uses data or values from","papers":1,"sentences":1,"quotes":[{"arxiv_id":"2402.17753","title":"Evaluating Very Long-Term Conversational Memory of LLM Agents","url":"/paper/2402.17753","section":"Motivating Analysis: Strategic Blindness","sentence":"To empirically validate this diagnosis, we conduct a preliminary study on a subset of LoCoMo [ref] , focusing on reasoning-intensive queries by excluding adversarial samples.","intent_p":0.94,"contests_p":null}]},{"kind":"comparison","label":"Compares with","papers":4,"sentences":4,"quotes":[{"arxiv_id":"2504.19413","title":"Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memory","url":"/paper/2504.19413","section":"Experimental Setup","sentence":"We compare against two passive baselines: Full Text and Naive RAG [ref] , and four active memory systems: LangMem [ref] , Mem0 [ref] , A-Mem [ref] , and LightMem [ref] .","intent_p":0.97,"contests_p":null},{"arxiv_id":"2312.10997","title":"Retrieval-Augmented Generation for Large Language Models: A Survey","url":"/paper/2312.10997","section":"Experimental Setup","sentence":"We compare against two passive baselines: Full Text and Naive RAG [ref] , and four active memory systems: LangMem [ref] , Mem0 [ref] , A-Mem [ref] , and LightMem [ref] .","intent_p":0.97,"contests_p":null},{"arxiv_id":"2502.12110","title":"A-MEM: Agentic Memory for LLM Agents","url":"/paper/2502.12110","section":"Experimental Setup","sentence":"We compare against two passive baselines: Full Text and Naive RAG [ref] , and four active memory systems: LangMem [ref] , Mem0 [ref] , A-Mem [ref] , and LightMem [ref] .","intent_p":0.97,"contests_p":null},{"arxiv_id":"2510.18866","title":"LightMem: Lightweight and Efficient Memory-Augmented Generation","url":"/paper/2510.18866","section":"Experimental Setup","sentence":"We compare against two passive baselines: Full Text and Naive RAG [ref] , and four active memory systems: LangMem [ref] , Mem0 [ref] , A-Mem [ref] , and LightMem [ref] .","intent_p":0.97,"contests_p":null}]},{"kind":"background","label":"Cites as background","papers":4,"sentences":4,"quotes":[{"arxiv_id":"2502.12110","title":"A-MEM: Agentic Memory for LLM Agents","url":"/paper/2502.12110","section":"Related Work","sentence":"Prior work improves long-term memory from several directions, including memory architecture [ref] , memory organization and consolidation [ref] , and memory retrieval [ref] .","intent_p":0.96,"contests_p":null},{"arxiv_id":"2510.18866","title":"LightMem: Lightweight and Efficient Memory-Augmented Generation","url":"/paper/2510.18866","section":"Related Work","sentence":"Prior work improves long-term memory from several directions, including memory architecture [ref] , memory organization and consolidation [ref] , and memory retrieval [ref] .","intent_p":0.96,"contests_p":null},{"arxiv_id":"2512.20237","title":"Knowing What is Missing: Efficient Conversational Memory via Explicit Evidence-Gap Tracking","url":"/paper/2512.20237","section":"Related Work","sentence":"Prior work improves long-term memory from several directions, including memory architecture [ref] , memory organization and consolidation [ref] , and memory retrieval [ref] .","intent_p":0.96,"contests_p":null},{"arxiv_id":"2307.09288","title":"Llama 2: Open Foundation and Fine-Tuned Chat Models","url":"/paper/2307.09288","section":"Introduction","sentence":"Large language models (LLMs) [ref] are evolving from episodic chatbots into persistent agentic systems [ref] that execute complex workflows over days or weeks.","intent_p":0.99,"contests_p":null}]}],"unlabelled_sentences":18},"claims":[],"thresholds":{"intent_p":0.8,"supports_p":0.95,"contests_p":0.7},"method":"Citing sentences come from each paper's LaTeX source; every \\cite marks the sentence and the works it cites. Pith's classifier typed each sentence. A label is shown at intent probability 0.8 or more (0.95 for support), a contest at 0.7 or more. Measured accuracy against two blind model families is at https://pith.science/accuracy. The quoted sentences are the citing authors' own.","held_up":null,"_links":{"html":"https://pith.science/paper/2603.18718"}}