{"arxiv_id":"2608.11218","incoming":{"papers":0,"sentences":0,"groups":[],"unlabelled_sentences":0},"outgoing":{"papers":32,"sentences":42,"groups":[{"kind":"method_or_tool","label":"Uses the method or tool of","papers":2,"sentences":2,"quotes":[{"arxiv_id":"2505.09388","title":"Qwen3 Technical Report","url":"/paper/2505.09388","section":"Datasets and reproducibility","sentence":"The main results use Qwen3-1.7B [ref] with greedy decoding.","intent_p":0.97,"contests_p":null},{"arxiv_id":"2607.01308","title":"Cache Merging as a Convergent Replicated State for Multi-Agent Latent Reasoning","url":"/paper/2607.01308","section":"Store state and convergence","sentence":"To be read, the set is rendered into a single cache by placing each fragment's block at a canonical range of positions, which fixes its rotary (RoPE) coordinates; because that placement is a deterministic function of the set, the render is byte-identical across delivery orders, a construction we take from prior work on convergent cache merging and describe in Appendix [ref] [ref] .","intent_p":0.88,"contests_p":null}]},{"kind":"comparison","label":"Compares with","papers":2,"sentences":2,"quotes":[{"arxiv_id":"2005.11401","title":"Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks","url":"/paper/2005.11401","section":"Related work","sentence":"Retrieval-augmented generation [ref] and its latent-space successors [ref] fetch text or learned document representations selected for a known query; we address raw KV fragments produced query-blind, and we set out the honest trade this makes against text retrieval in Section [ref].","intent_p":0.93,"contests_p":null},{"arxiv_id":"2511.18659","title":null,"url":"/paper/2511.18659","section":"Related work","sentence":"Retrieval-augmented generation [ref] and its latent-space successors [ref] fetch text or learned document representations selected for a known query; we address raw KV fragments produced query-blind, and we set out the honest trade this makes against text retrieval in Section [ref].","intent_p":0.93,"contests_p":null}]},{"kind":"supports_this_paper","label":"Cites as support","papers":2,"sentences":2,"quotes":[{"arxiv_id":"2505.12514","title":null,"url":"/paper/2505.12514","section":"Related work","sentence":"Latent tokens transport and scaffold computation rather than perform it [ref] , and multi-hop composition is built by a model's own attention over a sequential chain rather than by colocating independent vectors [ref] .","intent_p":1.0,"contests_p":null},{"arxiv_id":"2512.21711","title":"Do Latent Tokens Think? A Causal and Adversarial Analysis of Chain-of-Continuous-Thought","url":"/paper/2512.21711","section":"Related work","sentence":"Latent tokens transport and scaffold computation rather than perform it [ref] , and multi-hop composition is built by a model's own attention over a sequential chain rather than by colocating independent vectors [ref] .","intent_p":1.0,"contests_p":null}]},{"kind":"background","label":"Cites as background","papers":18,"sentences":20,"quotes":[{"arxiv_id":"2602.03695","title":"Agent Primitives: Reusable Latent Building Blocks for Multi-Agent Systems","url":"/paper/2602.03695","section":"Introduction","sentence":"A growing line of work has language-model agents reason and communicate in latent space rather than in text: instead of exchanging natural-language messages, each agent encodes its input into the transformer's key–value (KV) cache and passes that latent state on [ref] .","intent_p":0.96,"contests_p":null},{"arxiv_id":"2511.20639","title":"Latent Collaboration in Multi-Agent Systems","url":"/paper/2511.20639","section":"Introduction","sentence":"A growing line of work has language-model agents reason and communicate in latent space rather than in text: instead of exchanging natural-language messages, each agent encodes its input into the transformer's key–value (KV) cache and passes that latent state on [ref] .","intent_p":0.96,"contests_p":null},{"arxiv_id":"2511.09149","title":"Enabling Agents to Communicate Entirely in Latent Space","url":"/paper/2511.09149","section":"Related work","sentence":"LatentMAS [ref] prepends one agent's KV cache to the next in sequence; Agent Primitives [ref] run parallel solvers whose caches a selector consumes; Interlat [ref] , cache-to-cache transfer [ref] , and thought communication [ref] pass latent messages between models, and RecursiveMAS [ref] loops hidden states through a trained link.","intent_p":0.94,"contests_p":null},{"arxiv_id":"2604.25917","title":"Recursive Multi-Agent Systems","url":"/paper/2604.25917","section":"Related work","sentence":"LatentMAS [ref] prepends one agent's KV cache to the next in sequence; Agent Primitives [ref] run parallel solvers whose caches a selector consumes; Interlat [ref] , cache-to-cache transfer [ref] , and thought communication [ref] pass latent messages between models, and RecursiveMAS [ref] loops hidden states through a trained link.","intent_p":0.94,"contests_p":null},{"arxiv_id":"2510.03215","title":null,"url":"/paper/2510.03215","section":"Related work","sentence":"LatentMAS [ref] prepends one agent's KV cache to the next in sequence; Agent Primitives [ref] run parallel solvers whose caches a selector consumes; Interlat [ref] , cache-to-cache transfer [ref] , and thought communication [ref] pass latent messages between models, and RecursiveMAS [ref] loops hidden states through a trained link.","intent_p":0.94,"contests_p":null}]}],"unlabelled_sentences":16},"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/2608.11218"}}