GrACE: A Generative Approach to Better Confidence Elicitation and Efficient Test-Time Scaling in Large Language Models
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914449782210560 |
|---|---|
| author | Zhang, Zhaohan Liu, Ziquan Patras, Ioannis |
| author_facet | Zhang, Zhaohan Liu, Ziquan Patras, Ioannis |
| contents | Assessing the reliability of Large Language Models (LLMs) by confidence elicitation is a prominent approach to AI safety in high-stakes applications, such as healthcare and finance. Existing methods either require expensive computational overhead or suffer from poor calibration, making them impractical and unreliable for real-world deployment. In this work, we propose GrACE, a Generative Approach to Confidence Elicitation that enables scalable and reliable confidence elicitation for LLMs. GrACE adopts a novel mechanism in which the model expresses confidence by the similarity between the last hidden state and the embedding of a special token appended to the vocabulary, in real-time. We fine-tune the model for calibrating the confidence with targets associated with accuracy. Extensive experiments show that the confidence produced by GrACE achieves the best discriminative capacity and calibration on open-ended generation tasks without resorting to additional sampling or an auxiliary model. Moreover, we propose two confidence-based strategies for test-time scaling with GrACE, which not only improve the accuracy of the final decision but also significantly reduce the number of required samples, highlighting its potential as a practical solution for deploying LLMs with reliable, on-the-fly confidence estimation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_09438 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | GrACE: A Generative Approach to Better Confidence Elicitation and Efficient Test-Time Scaling in Large Language Models Zhang, Zhaohan Liu, Ziquan Patras, Ioannis Computation and Language Assessing the reliability of Large Language Models (LLMs) by confidence elicitation is a prominent approach to AI safety in high-stakes applications, such as healthcare and finance. Existing methods either require expensive computational overhead or suffer from poor calibration, making them impractical and unreliable for real-world deployment. In this work, we propose GrACE, a Generative Approach to Confidence Elicitation that enables scalable and reliable confidence elicitation for LLMs. GrACE adopts a novel mechanism in which the model expresses confidence by the similarity between the last hidden state and the embedding of a special token appended to the vocabulary, in real-time. We fine-tune the model for calibrating the confidence with targets associated with accuracy. Extensive experiments show that the confidence produced by GrACE achieves the best discriminative capacity and calibration on open-ended generation tasks without resorting to additional sampling or an auxiliary model. Moreover, we propose two confidence-based strategies for test-time scaling with GrACE, which not only improve the accuracy of the final decision but also significantly reduce the number of required samples, highlighting its potential as a practical solution for deploying LLMs with reliable, on-the-fly confidence estimation. |
| title | GrACE: A Generative Approach to Better Confidence Elicitation and Efficient Test-Time Scaling in Large Language Models |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2509.09438 |