Adaptive Consensus in LLM Ensembles via Sequential Evidence Accumulation: Automatic Budget Identification and Calibrated Commit Signals
On the map: LLM reasoning → Test-time compute and efficient thinking
Abstract (arXiv)
Large Language Model ensembles improve reasoning accuracy, but only up to a performance boundary beyond which additional deliberation degrades accuracy. We introduce DASE (Deliberative Adaptive Stopping Ensemble), a stopping heuristic for iterative ensemble deliberation that commits early on genuine consensus and applies a global-frequency fallback on fragmented evidence. We make three contributions. (1) DASE produces a commit-type routing partition that generalises across benchmarks and is complementary to verbalized single-call confidence. On GPQA-Extended (N=546, 70B ensemble), the partition yields a 39.5 pp routing gap (right-wall 81.1% vs. left-wall 41.5%). On AIME 2010-2023 (N=261, 120B ensemble, 3 seeds), right-wall commits reach 98.3% accuracy vs. left-wall 72.8% (25.5 pp gap), statistically equivalent to Opus 4.6 Standard verbalized confidence at matched coverage (25.7 pp gap; bootstrap p=0.873); the two mechanisms disagree on 37% of routing assignments. (2) Adaptive stopping, not injection bandwidth, drives accuracy. On AIME-300, bandwidth accounts for only 0.3 pp (ns). On GPQA-Extended at the 120B tier, sparse injection (≈15 tokens/worker/round) achieves 70.9% with a 30.7 pp routing gap; dense injection (≈600 chars/worker/round) achieves 72.2% but with halved right-wall coverage and a narrower 18.9 pp gap. (3) Injection-based methods exhibit an inverted-U accuracy-vs-inference trajectory; this pattern is hypothesis-generating.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 8 of its 15 references that are in the map. First posted on arXiv 5 May 2026.
Builds on
8 of 15 references.
Closest ancestors · 2
Same direction, newest first
- 2024Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters6 Aug
- 2023Self-Refine: Iterative Refinement with Self-Feedback30 Mar
Other references · 6
- 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,601
- 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,748
- 2023Improving Factuality and Reasoning in Language Models through Multiagent Debatecited by 1,739
- 2023Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debatecited by 995
- 2023Semantic Uncertainty: Linguistic Invariances for Uncertainty Estimation in Natural Language Generationcited by 795
- 2023Just Ask for Calibration: Strategies for Eliciting Calibrated Confidence Scores from Language Models Fine-Tuned with Human Feedbackcited by 788
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2024Efficient Dynamic Ensembling for Multiple LLM Experts
- 2026Reliability-Aware Adaptive Self-Consistency for Efficient Sampling in LLM Reasoning
- 2026Reasoning Consensus: Structural Ensembling of LLM Reasoning via Weighted DAG Aggregation
- 2026A Single Revision Step Improves Token-Efficient LLM Reasoning
- 2026Dual-Dimensional Consistency: Balancing Budget and Quality in Adaptive Inference-Time Scaling
- 2025CGES: Confidence-Guided Early Stopping for Efficient and Accurate Self-Consistency
- 2025ConSol: Sequential Probability Ratio Testing to Find Consistent LLM Reasoning Paths Efficiently
- 2026Sequential Consensus for Multi-Agent LLM Debates: A Wald-SPRT compute governor with calibration-based failure detection