Large language models (LLMs) often produce fluent but incorrect answers with unwarranted confidence. A central limitation is that standard LLMs represent uncertainty through a single predictive distribution, conflating epistemic ignorance with genuine ambiguity. We introduce Credal Large Language Models (CLLMs): an ensemble of LoRA adapters induces a credal set whose lower and upper probabilities expose the spread of plausible predictive distributions rather than collapsing to a single softmax output. From this representation we derive two complementary commitment scores. Credal Token Commitment (CTC) is a token-space score that combines lower-bound support, credal width, and intersection entropy, computed without additional generation. Semantic Commitment Consistency (SCC) extends commitment to semantic space using sampled completions, with SCC-Gap measuring the mismatch between token-level and semantic-level support. We evaluate hallucination detection, calibration, selective prediction, and reasoning on Gemma-2-9B, Llama-3.1-8B, and Qwen2.5-7B across OpenBookQA, CoQA, TriviaQA, and ARC-Challenge. CLLM is the best method on QA accuracy at competitive expected calibration error, and CTC tracks the best hallucination AUROC within 1.5 pp on most settings without additional generation. On selective prediction at 80% coverage, CLLM with SCC reaches 99.0% accuracy on OpenBookQA, and on ARC-Challenge CLLM with Csem confidence achieves <= 0.6% ECE across the three backbones.
Shuting Xie, Nathaniel Lesperance, Graham W. Taylorcs.LG cs.AI
Large language models (LLMs) are increasingly used for scientific decision support, yet reliable confidence estimation remains difficult in black-box settings. We study uncertainty estimation for hierarchical taxonomic reasoning generated by a black-box LLM in a long-tailed biodiversity monitoring pipeline. Using proxy features extracted by an open-source tool LLM, we train lightweight supervised estimators with hierarchy-aware supervision to predict rank-wise correctness. Across three tool LLMs, the supervised estimators consistently outperform a token-likelihood baseline for micro discrimination and selective prediction under a single global rejection threshold, improving micro AUROC from 0.57 to 0.75--0.80. The best results are achieved by a rank-specific multi-head design (H3), suggesting that accounting for hierarchical output structure is important when a unified abstention rule is required. Our code is publicly available at https://github.com/uoguelph-mlrg/hierarchy-aware-llm-uq
Financial AI systems often train information extractors on one textual register and deploy them across filings, news, and user-generated content, while standard F1 scores do not indicate which predictions remain safe to automate when the input distribution changes. We study confidence estimation and selective prediction for financial named entity recognition (NER) on a three-tier stress test spanning SEC filings, financial news, and general-topic social media as an extreme out-of-domain condition. We evaluate a BERT tagger and LoRA-tuned Qwen2.5-0.5B/1.5B models using five inference-time confidence signals, three training seeds, and bootstrap intervals. Confidence rankings themselves change under distribution shift: whole-output probability is the strongest in-domain error detector but deteriorates out of domain, whereas entity-span probability and self-consistency are more robust; self-consistency is also better calibrated without post-hoc fitting. Abstention reduces sentence error from 34.3% to below 2% on the highest-confidence 40% of in-domain inputs and remains useful on financial news, but recovers no usefully large clean subset under the extreme social-media shift. These results motivate a staged deployment strategy that detects severe distribution shift upstream before applying prediction-level confidence gating.
Large language models serve heterogeneous populations structured by domain, topic difficulty, and linguistic style. Conformal risk control (CRC) gives rigorous marginal risk guarantees for selective prediction with abstention, but marginal guarantees do not imply per-group ones: a model can meet the population budget while systematically over-exposing subgroups to errors. Under mild shift in group composition, standard CRC violates the budget in up to 47% of trials. We propose HG-CRC (Hierarchical Group-Conditional CRC), a post-hoc calibration framework enforcing simultaneous risk guarantees across all nodes of a user-defined group hierarchy. It applies a Bonferroni correction over nodes and a leaf-first policy that uses the most specific applicable threshold, falling back to coarser nodes when a finer one is uncertified or rejects the example. It needs only a held-out calibration set, with no retraining. We evaluate on three models (Qwen3-4B, Llama-3.1-8B-Instruct, Gemma-3-4B) and two benchmarks (ARC Challenge, MMLU-Pro) across eight configurations probing IID generalization, heterogeneity, mixture/domain/prompt/difficulty shift, label noise, and quantization. Main result: HG-CRC reaches an empirical 0% violation rate and WGER=0 on ARC Challenge for high-accuracy models (Qwen3-4B, Llama-3.1-8B). At 500 bootstrap trials these zeros are empirical upper bounds (true rate up to 0.6%), not certified. Results are benchmark-specific: on MMLU-Pro these models abstain entirely or (Llama) retain WGER=0.014. Gemma-3-4B, poorly calibrated here, degrades gracefully by abstaining. Participation cost vs. global CRC is 22 to 37 points. Ablations show hierarchical depth clears the budget: removing difficulty level returns violations to about 11%. Bonferroni is needed for the theoretical guarantee, though its empirical effect matters only with many nodes.
Can a language model read the quality of its ongoing computation, and can an external intervention turn that readout into better outcomes? We test both questions in a frozen 2.6B looped transformer, Ouro-RLTT. On GSM8K, a strict pre-answer probe excludes the answer region and gold value yet predicts success: hidden states plus length/log-probability features reach AUROC 0.797 versus 0.731 for those surface features alone (increment +0.066; task-clustered 95% CI [+0.021,+0.112]; 170 tasks). On Horizon Logic, a prospectively extended task-disjoint study gives an increment of +0.111 (CI [+0.056,+0.169]), independently replicated on the new cohort (+0.095) and robust to an adversarial malformed-sibling shortcut. Recurrence also moves candidate-quality readability to progressively earlier physical depth; the trend replicates across the Ouro family and qualitatively in out-of-family Huginn, although their transfer geometry differs. The readout converts into validated decision-level gains. Hidden-state-based scores improve risk-coverage over shortcut-only scores in four sealed selective-prediction arms, and terminal selection beats matched random even when every candidate is well formed (27/32 correct selections versus 64.8% expected; p = 0.0086). Generative control does not convert: directional steering is negative, a branch screen is bounded, and exact-compute loop allocation and minimal LoRA direction-binding detect no gain. These tests run through bit-exact branch/carry/prune machinery over Ouro's 192-slot recurrent cache, including a suffix-recompute splice saving up to 88% of per-branch layer passes. We call this decision-usable but not generatively controllable property operational proto-introspection. All load-bearing values use source-item-disjoint splits and antisymmetrized pairwise evaluation.
Evaluating uncertainty in AI-generated SQL queries requires estimating whether a query is correct, where correct means it executes to the same result as a human-written reference. We study which signals predict correctness on hard multi-table text-to-SQL, using AUROC to measure how well each ranks correct queries above incorrect ones. On BIRD and Spider, black-box signals such as string, structural, and execution self-consistency, a schema-relevance score, and query executability all fall between about 0.61 and 0.68 AUROC, with string self-consistency strongest at 0.675; white-box log-probability is similar (0.67). The signals that move past this ceiling are verification-based: an LLM judge scores from 0.72 (GPT-4o-mini) to 0.78 (Claude). Judges from different providers make different errors, so a two-provider ensemble reaches 0.82 AUROC with a well-calibrated probability (expected calibration error 0.03) and supports useful abstention frontiers (for example, answering 27% of questions at 24% selective risk) where self-consistency offers no valid low-risk subset. The pattern holds across two benchmarks, two generators, and two judge providers. We also ask whether a verifier can be trained. Fine-tuned verifiers, both encoder and generative, reach about 0.77 to 0.79 AUROC in-distribution but fall to about 0.66 on unseen schemas; scaling to 7B, adding schema diversity, distilling a strong judge's rationales, and cross-benchmark training all fail to close that gap. Cross-schema transfer appears to track model scale and reasoning rather than fine-tuning. In practice, correctness uncertainty for text-to-SQL lives in reasoning-based signals: a fine-tuned verifier is a good in-domain tool, but a verifier that generalizes across schemas currently means a large frozen reasoning model.
Andrea Alfarano, Andrea Bacciu, Saab Mansour +2cs.CL cs.AI
Uncertainty estimation (UE) enables LLM-powered systems to recognize when to abstain, yet existing research has predominantly focused on English. We present the first large-scale evaluation of UE methods across 22 languages, spanning high-, mid-, and low-resource settings. Using two human-curated Q\&A datasets, we compare open and closed box UE methods (nine in total) across different model sizes and architectures while eliciting long-form reasoning, avoiding LLM-as-a-judge and embedding-based scoring, which can introduce evaluation noise. We report three main actionable findings. First, we find that prompting models to reason in English while keeping questions in low-resource languages substantially improves UE performance, suggesting that comprehension of low-resource languages is largely intact, and that the reliability bottleneck lies in generation rather than understanding. Second, prompting models to reason in English closes the UE performance gap between low and high-resource languages, demonstrating that generation language matters more than the question language. Third, the choice of UE method should depend on model scale: at smaller scales, open-box probability-based methods outperform alternatives; at larger scales, closed-box self-verbalized uncertainty becomes superior. Finally, we provide an analysis of threshold selection for selective prediction, offering guidance on calibrating abstention in multilingual settings.
Large language models (LLMs) are increasingly deployed as critical decision-making components in high-stakes real-world AI systems, rendering LLM reliability a foremost practical concern. In this paper, we focus on enhancing LLM reliability through selective prediction (SP), a strategy that allows an LLM to only predict for inputs where it is likely to be correct (i.e., coverage) and hence reduce the error rate (i.e., risk) on that portion of inputs -- flagging the remaining inputs for future human discretion. In other words, SP improves LLM reliability by balancing the risk-coverage trade-off and enabling seamless human-AI collaboration. To integrate SP into LLMs, we focus on the LLM post-training alignment stage and propose to align LLMs with SP performance metrics, in contrast with existing LLM alignment methods that focus primarily on correctness or calibration metrics. Specifically, we propose a novel alignment framework, Reinforcement Learning for Selection Reward (RLSR), which targets the area under the risk-coverage curve (AURC) -- a popular SP performance metric -- as its alignment objective. RLSR achieves substantially better risk-coverage trade-off compared to multiple alignment baselines on both in-domain and out-of-domain tasks.
In high-stakes document processing pipelines, including financial reconciliation, compliance verification, and procurement automation, an LLM extraction that is silently wrong is more dangerous than one that is visibly absent. The central challenge is not extraction accuracy alone but reliable confidence estimation: knowing, field by field, whether an extraction can be trusted for automation or deferred to human review. Token-level log-probabilities, verbalized confidence, and multi-sample self-consistency all collapse toward all-positive behaviour at practical thresholds, offering no reliable separation between trustworthy and untrustworthy extractions. We present ExtractConf, a cross-domain, field-agnostic confidence engine that grounds confidence estimation in two structurally different readings of the same document. A field-guided Hunter call extracts each field under schema-slot completion pressure; a document-guided Mapper call scans holistically and surfaces values grounded in document content. This asymmetry yields different failure modes: Hunter hallucinates values for absent fields, while Mapper misses visually non-salient ones. Their disagreement is independently informative. ExtractConf fuses cross-call disagreement, LLM-internal uncertainty, OCR, image quality, and spatial layout into a classifier requiring no domain-specific rules or retraining. On DocILE (55-field invoices, 26% failure rate), it achieves 0.928 ROC AUC and reduces selective prediction risk by 70% over logprob-mean. At 80% coverage, accuracy reaches 99.1%, enabling a practical human-in-the-loop workflow. Zero-shot transfer to CORD receipts achieves 0.858 AUC; lightweight Lasso recalibration reduces ECE by 89% and Brier by 43%, confirming the signals generalise across document domains.
Large language models (LLMs) are increasingly deployed as black-box classifiers in pipelines that automate confident decisions and route uncertain ones to human review. Such selective prediction needs a confidence score that an operator can threshold at a chosen risk level. Prior work asks whether LLM confidence is well calibrated or well ranked; we ask a complementary, deployment-oriented question that has been largely overlooked: at what resolution can the score be thresholded? We call the answer the score granularity gap. Through a controlled comparison of seven ways to build a confidence score, from a single verbalized number, to token probabilities, to querying the model many times and combining the answers, across 25 model-dataset pairs (9 LLMs, 3 benchmarks), we find that single-shot verbalized confidence, once correctly converted to a class probability, ranks cases surprisingly well, yet takes only a handful of distinct values. It therefore offers an operator only a few coarse thresholds, no matter how well it ranks. We show which constructions widen this gap, at what inference cost, and with what effect on ranking, notably that multi-query aggregation helps weak models but can degrade already-strong ones. We translate these trade-offs into concrete deployment guidance.
Detecting LLM reasoning failures at inference time without ground-truth labels has motivated a wide range of confidence baselines, including self-consistency, semantic entropy, and P(True), built on within-question sampling and self-evaluation. Operad theory, the formalism for systems built by iterated substitution, suggests a complementary diagnostic: a model's direct answer to a compositional query should agree with the answer it produces by composing a stated decomposition of the same query. We instantiate this idea as operadic consistency (OC), a per-question signal. Across twelve instruction-tuned LLMs (4B to 671B parameters, open-weights and closed-source) on four multi-hop QA datasets, OC is strongly correlated with accuracy on every dataset (Pearson $r \in [0.86, 0.94]$, all $p \leq 0.0004$), and is the only signal we evaluate with $r \geq 0.85$ uniformly across all four datasets. Chain-of-thought self-consistency (CoT-SC; Wang et al., 2023) matches OC on HotpotQA and DROP ($r = 0.93, 0.87$) but drops to $r \approx 0.45$ on MuSiQue and StrategyQA. At the per-question level, OC contributes information beyond CoT-SC and semantic entropy on every dataset (cluster-robust $p \leq 10^{-16}$ for the OC coefficient), and the conclusion is robust to additionally controlling for constructed decomposition-aware baselines ($p \leq 10^{-13}$). The same signal yields selective-prediction improvements (accuracy at fixed coverage) over a tuned CoT-SC baseline at the equal-cost $K = 3$ budget (AUARC lifts of +0.086 to +0.096 and AUROC lifts of +0.092 to +0.164; 95% CIs exclude zero on every cell). On five frontier thinking models, where the decomposition is extracted from the model's own chain of thought, the same equal-cost comparison gives positive selective-prediction point-estimate lift on all 16 (dataset, budget, metric) cells tested, with 95% CIs excluding zero on 12 of the 16.
Hallucination detection in large language and vision-language models is increasingly framed as selective prediction, where a detector assigns a confidence score and abstains when confidence is low. Unsupervised sampling detectors (Semantic Entropy) avoid labels but plateau in quality, while supervised probes attain stronger in-distribution scores yet degrade sharply when calibration labels are scarce. We recover the response manifold of an LLM as the density ridge of a kernel density estimate built on a six-dimensional kinematic feature map of hidden state generation trajectories. A test generation is scored by the negated Euclidean distance from its projected feature point to the nearest ridge vertex, yielding a low-dimensional geometric skeleton of the stochastic output distribution. We evaluate against Semantic Entropy, topological methods, and log-probability on six QA benchmarks (HaluEval-QA, TriviaQA, GSM8K, POPE, ScienceQA, A-OKVQA) using eight text and vision LLMs in a deliberately label-scarce protocol ($n_{\text{cal}}{=}200$ queries, $N{=}5$ generations). Our ridge-based score beats on AUROC with 5-20 points gain, while demonstrating tempered degradation under calibration-label scarcity.
Hadi Mohammadi, Shihan Wang, Masoume M. Raeissi +1cs.CL
The detection of online sexism remains an open problem. Sexism detection is inherently subjective, yet most existing systems reduce multi-annotator labels to a single majority decision and treat all instances uniformly. This ignores two informative signals: annotator agreement and model uncertainty. We propose RA-DPO (Reliability-Aware Direct Preference Optimization), which integrates annotator agreement, model confidence, and a token-level uncertainty signal into a single reliability score. RA-DPO uses this score to select high-value preference pairs during training and to support inference-time abstention, which allows the model to trade coverage for accuracy. We evaluate RA-DPO on 6,920 multilingual posts from EXIST 2023, fine-tune OpenAI gpt-4o base via DPO, and validate on two open-weight 3B models (Llama, Qwen). Results show that training on the top 30% most reliable pairs matches full-data DPO, which indicates that reliability-aware selection can reduce training cost without sacrificing performance. At inference, selective prediction reaches 96.2% accuracy at 50% coverage in the true-agreement setting and 88.7% in the deployable predicted-agreement setting, both exceeding the 85.3% no-agreement baseline. These results suggest that accounting for annotation uncertainty is beneficial for both efficient training and reliable deployment in subjective classification.