CGES: Confidence-Guided Early Stopping for Efficient and Accurate Self-Consistency

• URL: http://arxiv.org/abs/2511.02603v1
• Date: Tue, 04 Nov 2025 14:25:54 GMT
• Title: CGES: Confidence-Guided Early Stopping for Efficient and Accurate Self-Consistency
• Abstract summary: Large language models (LLMs) are often queried multiple times at test time, with predictions aggregated by majority vote.<n>We introduce Confidence-Guided Early Stopping (CGES), a Bayesian framework that forms posteriors over candidate answers.<n>CGES adaptively halts sampling once the posterior mass of a candidate exceeds a threshold.
• Score: 3.2807118426157493
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large language models (LLMs) are often queried multiple times at test time, with predictions aggregated by majority vote. While effective, this self-consistency strategy (arXiv:2203.11171) requires a fixed number of calls and can fail when the correct answer is rare. We introduce Confidence-Guided Early Stopping (CGES), a Bayesian framework that forms posteriors over candidate answers using scalar confidence signals derived from token probabilities or reward models. CGES adaptively halts sampling once the posterior mass of a candidate exceeds a threshold. We provide theoretical guarantees for both perfectly calibrated confidences and realistic noisy confidence signals. Across five reasoning benchmarks, CGES reduces the average number of model calls by about 69 percent (for example, from 16.0 to 4.9) while matching the accuracy of self-consistency within 0.06 percentage points.

Related papers

• SVR: Self-Verifying Refinement via Joint Verdict-Confidence Reinforcement Learning for Adaptive Test-Time Compute [62.3458279176813]
  Self-Verifying Refinement is an oracle-free multi-turn reinforcement learning framework.<n>It learns to use self-verification as a compute-control policy.<n>It achieves a macro-average accuracy of 0.563 with only 2.99 inference turns on average.
  arXiv  Detail & Related papers  (2026-07-30T16:20:58Z)
• CALIBER: Calibrating Confidence Before and After Reasoning in Language Models [16.376895132982938]
  We introduce CALIBER (Calibration Before and After Reasoning), which elicits both estimates each with the target matched to its information state.<n>Under this unified protocol, CALIBER reduces Expected Error (ECE) by 52.5% over the strongest single-confidence baseline on BigMathDigits.<n>On a larger 30B model, CALIBER achieves the best ECE on BigMathDigits while remaining competitive in Brier score and AUROC.
  arXiv  Detail & Related papers  (2026-06-23T08:03:20Z)
• Boosting Self-Consistency with Ranking [56.38798757709555]
  Self-consistency improves large language models by sampling multiple reasoning paths and selecting the most frequent answer.<n>We address this limitation with Ranking-Improved Self-Consistency (RISC), which reformulates answer selection in self-consistency as a ranking problem.
  arXiv  Detail & Related papers  (2026-06-03T16:12:30Z)
• Understanding and Mitigating Premature Confidence for Better LLM Reasoning [76.16007941549857]
  Long chains of thought (CoT) from current language models frequently contain logical gaps and unjustified leaps.<n>We find such a signal in how the model's confidence evolves during reasoning.<n>We exploit this in progressive confidence shaping, a reinforcement learning objective that trains models to update their confidence as they reason rather than commit early.
  arXiv  Detail & Related papers  (2026-05-23T04:42:45Z)
• Pause and Reflect: Conformal Aggregation for Chain-of-Thought Reasoning [8.024041325202612]
  Chain-of-thought (CoT) reasoning with self-consistency improves performance by aggregating multiple sampled reasoning paths.<n>We introduce a conformal procedure for CoT reasoning that directly addresses aggregation uncertainty.<n>Our approach replaces majority voting with weighted score aggregation over reasoning paths and calibrates an abstention rule using conformal risk control.
  arXiv  Detail & Related papers  (2026-05-13T20:33:59Z)
• What Single-Prompt Accuracy Misses: A Multi-Variant Reliability Audit of Language Models [0.0]
  Single-prompt accuracy is the dominant way to benchmark language models, but it can miss reliability failures that matter.<n>We evaluate a 15-model open-weight corpus, with the main reliability analyses focused on 10 instruct models across five classification and reasoning benchmarks.<n>We find three broad results. First, evaluation design can materially change the conclusion.<n>Second, confidence signals are fragile. On MMLU-Pro, every primary model verbally reports confidence substantially above both its accuracy and its token-probability confidence on the same rows, and verbal parse rate can collapse for a single model on a single prompt variant.
  arXiv  Detail & Related papers  (2026-05-03T20:05:08Z)
• Confidence Before Answering: A Paradigm Shift for Efficient LLM Uncertainty Estimation [47.91529693614168]
  Existing methods are predominantly answer-first, producing confidence only after generating an answer.<n>We study a confidence-first paradigm, where the model outputs its confidence before answering, interpreting this score as the model's probability of answering the question correctly.<n>We propose CoCA, a reinforcement learning framework that jointly optimize confidence calibration and answer accuracy via segmented credit assignment.
  arXiv  Detail & Related papers  (2026-03-06T04:03:13Z)
• CoRefine: Confidence-Guided Self-Refinement for Adaptive Test-Time Compute [10.548368675645403]
  CoRefine is a confidence-guided self-refinement method that achieves competitive accuracy using a fraction of the tokens.<n>The controller consumes full-trace confidence to decide whether to halt, re-examine, or try a different approach.<n>We extend this to CoRefine-Tree, a hybrid sequential-parallel variant that adaptively balances exploration and exploitation.
  arXiv  Detail & Related papers  (2026-02-09T17:44:41Z)
• CARE What Fails: Contrastive Anchored-REflection for Verifiable Multimodal [84.71254539482369]
  Group-relative reinforcement learning with verifiable rewards (RLVR) often wastes the most informative data it already has the failures.<n>We present CARE, a failure-centric post-training framework for multimodal reasoning that turns errors into supervision.<n> CARE improves accuracy and training smoothness while explicitly increasing the share of learning signal that comes from failures.
  arXiv  Detail & Related papers  (2025-12-22T16:34:21Z)
• Stopping Rules for Stochastic Gradient Descent via Anytime-Valid Confidence Sequences [51.56484100374058]
  We study stopping rules for gradient descent (SGD) for convex optimization.<n>We develop an anytime-valid, data-dependent upper confidence sequence for the weighted average suboptimality of projected SGD.<n>These are the first rigorous, time-uniform performance guarantees and finitetime $varepsilon$-optimality certificates.
  arXiv  Detail & Related papers  (2025-12-15T09:26:45Z)
• LYNX: Learning Dynamic Exits for Confidence-Controlled Reasoning [15.597220136913258]
  LYNX is an online early-exit mechanism that turns a model's own hidden-state awareness into confidence-controlled stopping decisions.<n>We train and calibrate this probe once on a generic mathematical corpus and reuse it unchanged across benchmarks, decoding temperatures, and even non-mathematical tasks.
  arXiv  Detail & Related papers  (2025-12-05T00:04:42Z)
• LLMs are Overconfident: Evaluating Confidence Interval Calibration with FermiEval [0.3499870393443268]
  Large language models (LLMs) excel at numerical estimation but struggle to correctly quantify uncertainty.<n>We study how well LLMs construct confidence intervals around their own answers and find that they are systematically overconfident.
  arXiv  Detail & Related papers  (2025-10-30T20:49:41Z)
• Sample Smart, Not Hard: Correctness-First Decoding for Better Reasoning in LLMs [72.82403830490084]
  We argue that the decoding rule should be calibrated by correctness, not confidence alone.<n>We propose simple strategies that achieve this goal: Greedy-Threshold makes sampling greedy at very low confidence steps.<n>Together, our findings challenge prevailings about decoding under uncertainty and show gains across math and general reasoning benchmarks.
  arXiv  Detail & Related papers  (2025-10-07T14:46:12Z)
• ConfRAG: Confidence-Guided Retrieval-Augmenting Generation [41.78313747240249]
  We introduce ConfQA, a fine-tuning strategy that reduces hallucination rates from 20-40% to below 5% across multiple factuality benchmarks.<n>We propose ConfRAG, a triggering strategy that invokes RAG only when the model responses with unsure.<n>This framework achieves accuracy above 95% in ideal case while reducing unnecessary external retrievals by over 30%.
  arXiv  Detail & Related papers  (2025-06-08T22:51:46Z)
• Efficient Test-Time Scaling via Self-Calibration [18.32718448734639]
  Best-of-N sampling and Self-Consistency with majority voting are simple and effective, but require a fixed number of sampling responses for each query.<n>This could result in wasted computation for simpler questions and insufficient exploration for more challenging ones.<n>We argue that model confidence of responses can be used for improving the efficiency of test-time scaling.
  arXiv  Detail & Related papers  (2025-02-25T00:21:14Z)
• Confidence Improves Self-Consistency in LLMs [17.280967928501678]
  We introduce Confidence-Informed Self-Consistency (CISC)<n>CISC performs a weighted majority vote based on confidence scores obtained directly from the model.<n>When tested on nine models and four datasets, CISC outperforms self-consistency in nearly all configurations.
  arXiv  Detail & Related papers  (2025-02-10T08:10:29Z)
• Llamas Know What GPTs Don't Show: Surrogate Models for Confidence Estimation [70.27452774899189]
  Large language models (LLMs) should signal low confidence on examples where they are incorrect, instead of misleading the user. As of November 2023, state-of-the-art LLMs do not provide access to these probabilities. Our best method composing linguistic confidences and surrogate model probabilities gives state-of-the-art confidence estimates on all 12 datasets.
  arXiv  Detail & Related papers  (2023-11-15T11:27:44Z)
• When Does Confidence-Based Cascade Deferral Suffice? [69.28314307469381]
  Cascades are a classical strategy to enable inference cost to vary adaptively across samples. A deferral rule determines whether to invoke the next classifier in the sequence, or to terminate prediction. Despite being oblivious to the structure of the cascade, confidence-based deferral often works remarkably well in practice.
  arXiv  Detail & Related papers  (2023-07-06T04:13:57Z)
• Just Ask for Calibration: Strategies for Eliciting Calibrated Confidence Scores from Language Models Fine-Tuned with Human Feedback [91.22679548111127]
  A trustworthy real-world prediction system should produce well-calibrated confidence scores. We show that verbalized confidences emitted as output tokens are typically better-calibrated than the model's conditional probabilities.
  arXiv  Detail & Related papers  (2023-05-24T10:12:33Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.