RCWT: Measuring Task-Budget Displacement from Coordination Content in LLM Calls
Quick Answer
This paper shows that The RCWT protocol measures task-budget displacement in LLM calls, revealing that coordination tokens significantly impact performance.
Quick Take
In tests with GPT-4.1-mini, Claude Haiku 4.5, and Gemini 2.5 Flash, models maintained accuracy up to a 95% coordination ratio, indicating that task-budget displacement is a key factor rather than coordination volume alone causing semantic interference.
Key Points
- RCWT varies coordination content while controlling total budget and task family.
- Three commercial models showed baseline performance degradation with increased coordination tokens.
- Models maintained accuracy with intact task evidence despite high coordination ratios.
- RCWT serves as a measurement primitive for context-allocation budgeting.
- Findings suggest task-budget displacement impacts performance significantly.
Paper Resources
📖 Reader Mode
~2 min readAbstract:Multi-agent and memory-augmented LLM systems often place coordination content, shared state, prior discussion, tool outputs, summaries, and role instructions, inside the same finite prompt used for the current task. This creates a practical allocation problem: every token spent on coordination is unavailable to task instructions or evidence when a call is assembled under a fixed context budget. We introduce the Roundtable Context Window Test (RCWT), a controlled protocol for measuring this task-budget displacement effect. RCWT varies coordination content while controlling total budget, position order, task family, and scoring. In the main context-dependent recall task at $W=4096$, three commercial models remain near baseline through moderate overhead and then degrade sharply once residual reference evidence falls to a few hundred tokens. Window-scaling summaries are consistent with a task-specific residual-budget interpretation rather than a fixed percentage threshold, but we treat this as descriptive evidence rather than a universal law. To test whether the fixed-budget cliff persists when task evidence remains intact, we add an intact-task ablation: the full task/reference block is kept present while coordination tokens increase by expanding total prompt length. In that setting, all tested calls return every scored field correctly across GPT-4.1-mini, Claude Haiku 4.5, and Gemini 2.5 Flash up to a 95\% coordination ratio. This ablation narrows the claim: the main RCWT cliff is best read as task-budget displacement, not as proof that coordination volume alone causes semantic interference in the original open-ended task. RCWT is therefore a measurement primitive for context-allocation budgeting, not a complete theory of multi-agent benefit or session-level coordination.
| Comments: | 10 pages, 1 figure |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA) |
| Cite as: | arXiv:2607.12216 [cs.CL] |
| (or arXiv:2607.12216v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2607.12216 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Rodrigo Da Motta Cabral De Carvalho [view email]
[v1]
Mon, 13 Jul 2026 23:31:44 UTC (82 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CL
See more →TriAgent: Divergence-Aware Committees for Cost-Efficient Financial Sentiment Analysis
TriAgent introduces a cost-efficient multi-agent system for financial sentiment analysis, combining VADER, FinBERT, and Qwen2.5. It achieves an F1 score of ~0.87 with significant savings of $9.3M/year at a 10M-user scale compared to GPT-4o-mini, while also detecting hallucinations with an AUC of 0.90.