When Does Latent Communication Pay? A Causal Audit of Relayed KV Caches in Multi-Agent LLMs
Multi-agent LLM systems relay key-value caches instead of text and credit their gains to exchanged "latent thoughts". That credit is a claim about which example's cache is relayed, not merely that one is. We audit it causally in released systems. The cache is replaced with deranged (mismatched-example), zeroed, and moment-matched random counterparts, under two regimes defined by whether the receiver needs the sender's private information. Where it does, the battery reads ceiling: 100% against 23-25% for answer-irrelevant relays on the primary backbone, a contrast replicated across three families, five checkpoints, and a prose document-QA surface. Where it does not, a pre-registered five-seed protocol establishes equivalence within 2.8 points, a margin anchored to the audited system's reported gain, under Holm-corrected TOST on GSM8K and ARC-Challenge across three Qwen3 scales and on MedQA at 8B (one cell shows a small detected advantage inside the margin); a second family shows no detected advantage. A large cache effect need not be a pairing effect. In one natural cell, zeroing the relay costs 14.7 points; a mismatched cache, 0.4. Nor is need sufficient: under the same test, delivered channels span ceiling (LatentMAS's native relay), partial (KVComm's layer subset), and no detected example-specific transfer (C2C's released projector). Benchmark deltas do not by themselves establish latent-thought transmission; establishing it takes a mismatched-cache audit, which we release.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
Do Latent Channels Actually Communicate? A Causal Audit of Latent Multi-Agent LLM
Latent communication in large language model (LLM)-based multi-agent systems (MAS) transmits continuous internal representations instead of text, but greater representational capacity does not establish that the receiver…
RAudit: A Blind Auditing Protocol for Large Language Model Reasoning
Inference-time scaling can amplify reasoning pathologies: sycophancy, rung collapse, and premature certainty. We present RAudit, a diagnostic protocol for auditing LLM reasoning without ground truth access. The key const…
Mathematical ReasoningReplaySCM: A Benchmark for Executable Causal Mechanism Induction from Interventions
Most causal benchmarks for language models score local answers or graph structure. We introduce ReplaySCM, a 1,300 item benchmark for executable causal mechanism induction from finite interventional evidence. Each item c…
What Does a Meow Mean? In Search of Intuitively Understandable Communication by a Nonverbal Companion Robot
Older adults living alone have a number of challenges, and robots can help with some of them--by providing reminders, initiating activity, or offering comfort. As part of developing a cat robot with limited assistive fun…
Bypassing the Rationale: Causal Auditing of Implicit Reasoning in Language Models
Chain-of-thought (CoT) prompting is widely used as a reasoning aid and is often treated as a transparency mechanism. Yet behavioral gains under CoT do not imply that the model's internal computation causally depends on t…