dorsal/arxiv
View SchemaHow Reliable are Confidence Estimators for Large Reasoning Models? A Systematic Benchmark on High-Stakes Domains
| Authors | Reza Khanmohammadi, Erfan Miahi, Simerjot Kaur, Ivan Brugere, Charese H. Smiley, Kundan Thind, Mohammad M. Ghassemi |
|---|---|
| Categories | |
| ArXiv ID | 2601.08134vv2 |
| URL | https://arxiv.org/abs/2601.08134 |
| License | http://creativecommons.org/licenses/by/4.0/ |
Abstract
The miscalibration of Large Reasoning Models (LRMs) undermines their reliability in high-stakes domains, necessitating methods to accurately estimate the confidence of their long-form, multi-step outputs. To address this gap, we introduce the Reasoning Model Confidence estimation Benchmark (RMCB), a public resource of 347,496 reasoning traces from six popular LRMs across different architectural families. The benchmark is constructed from a diverse suite of datasets spanning high-stakes domains, including clinical, financial, legal, and mathematical reasoning, alongside complex general reasoning benchmarks, with correctness annotations provided for all samples. Using RMCB, we conduct a large-scale empirical evaluation of over ten distinct representation-based methods, spanning sequential, graph-based, and text-based architectures. Our central finding is a persistent trade-off between discrimination (AUROC) and calibration (ECE): text-based encoders achieve the best AUROC (0.672), while structurally-aware models yield the best ECE (0.148), with no single method dominating both. Furthermore, we find that increased architectural complexity does not reliably outperform simpler sequential baselines, suggesting a performance ceiling for methods relying solely on chunk-level hidden states. This work provides the most comprehensive benchmark for this task to date, establishing rigorous baselines and demonstrating the limitations of current representation-based paradigms.
{
"annotation_id": "4250f48a-06a0-4a8d-85a5-4dc76ab33387",
"date_created": "2026-02-17T05:53:15.942000Z",
"date_modified": "2026-02-17T05:53:15.942000Z",
"file_hash": "3036a26d49521cdb97c097dcbd3be3d25ef9170e6bcbc944888b9acbd4ceec08",
"private": false,
"record": {
"abstract": "The miscalibration of Large Reasoning Models (LRMs) undermines their reliability in high-stakes domains, necessitating methods to accurately estimate the confidence of their long-form, multi-step outputs. To address this gap, we introduce the Reasoning Model Confidence estimation Benchmark (RMCB), a public resource of 347,496 reasoning traces from six popular LRMs across different architectural families. The benchmark is constructed from a diverse suite of datasets spanning high-stakes domains, including clinical, financial, legal, and mathematical reasoning, alongside complex general reasoning benchmarks, with correctness annotations provided for all samples. Using RMCB, we conduct a large-scale empirical evaluation of over ten distinct representation-based methods, spanning sequential, graph-based, and text-based architectures. Our central finding is a persistent trade-off between discrimination (AUROC) and calibration (ECE): text-based encoders achieve the best AUROC (0.672), while structurally-aware models yield the best ECE (0.148), with no single method dominating both. Furthermore, we find that increased architectural complexity does not reliably outperform simpler sequential baselines, suggesting a performance ceiling for methods relying solely on chunk-level hidden states. This work provides the most comprehensive benchmark for this task to date, establishing rigorous baselines and demonstrating the limitations of current representation-based paradigms.",
"arxiv_id": "2601.08134",
"authors": [
"Reza Khanmohammadi",
"Erfan Miahi",
"Simerjot Kaur",
"Ivan Brugere",
"Charese H. Smiley",
"Kundan Thind",
"Mohammad M. Ghassemi"
],
"categories": [
"cs.CL"
],
"license": "http://creativecommons.org/licenses/by/4.0/",
"title": "How Reliable are Confidence Estimators for Large Reasoning Models? A Systematic Benchmark on High-Stakes Domains",
"url": "https://arxiv.org/abs/2601.08134",
"version": "v2"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "ef9ae321-ce29-467e-8e42-0b1a56c9ced9",
"id": "arXiv Dataset",
"type": "Model",
"variant": "snapshot-2026-01-17",
"version": "0.1.0"
},
"user_id": 1000002
}