paper-with-me

홈 › Papers

SAFER: Risk-Constrained Sample-then-Filter in Large Language Models

2025-10-11 · Qingni Wang, Yue Fan, Xin Eric Wang arxiv

As large language models (LLMs) are increasingly deployed in risk-sensitive applications such as real-world open-ended question answering (QA), ensuring the trustworthiness of their outputs has become critical. Existing selective conformal prediction (SCP) methods provide statistical guarantees by constructing prediction sets with a constrained miscoverage rate for correct answers. However, prior works unrealistically assume that admissible answers for all instances can be obtained via finite sampling, even for open-ended QA scenarios that lack a fixed and finite solution space. To address this, we introduce a two-stage risk control framework comprising abstention-aware sampling and conformalized filtering (SAFER). Firstly, on a held-out calibration set, SAFER calibrates a sampling budget within the maximum sampling cap, using the Clopper-Pearson exact method at a user-desired risk level (i.e., the maximum allowable miscoverage rate of the sampling sets). If the risk level cannot be satisfied within the cap, we abstain; otherwise, the calibrated sampling budget becomes the minimum requirements at test time. Then, we employ calibration instances where correct answers are attainable under the calibrated budget and apply the conformal risk control method to determine a statistically valid uncertainty threshold, which filters unreliable distractors from the candidate set for each test data point. In this stage, SAFER introduces an additional risk level to guide the calculation of the threshold, thereby controlling the risk of correct answers being excluded. Furthermore, we show that SAFER is compatible with various task-specific admission criteria and calibration-test split ratios, highlighting its robustness and high data efficiency.

📄 PDF Abstract BibTeX arXiv:2510.10193

Code (0)

등록된 구현이 없습니다.

Tasks

Question Answering

Similar Papers 제목 키워드 기반

Towards Safer Pretraining: Analyzing and Filtering Harmful Content in Webscale datasets for Responsible LLMs

2025-05-04 · Sai Krishna Mendu, Harish Yenala, Aditi Gulati, Shanu Kumar 외

Large language models (LLMs) have become integral to various real-world applications, leveraging massive, web-sourced datasets like Common Crawl, C4, and FineWeb for pretraining. While these datasets provide linguistic d…

MisinformationText Generation

MultiRisk: Multiple Risk Control via Iterative Score Thresholding

2025-12-31 · Sunay Joshi, Yan Sun, Hamed Hassani, Edgar Dobriban arxiv

As generative AI systems are increasingly deployed in real-world applications, regulating multiple dimensions of model behavior has become essential. We focus on test-time filtering: a lightweight mechanism for behavior …

SaFeR-VLM: Toward Safety-aware Fine-grained Reasoning in Multimodal Models

2025-10-08 · Huahui Yi, Kun Wang, Qiankun Li, Miao Yu 외 arxiv

Multimodal Large Reasoning Models (MLRMs) demonstrate impressive cross-modal reasoning but often amplify safety risks under adversarial or unsafe prompts, a phenomenon we call the \textit{Reasoning Tax}. Existing defense…

Reinforcement LearningMultimodal Reasoning

SafeRelBench: A Spatial-Relation-Aware Benchmark for Process-Level Safety in VLM-Driven Embodied Agents

2026-07-16 · Huaigang Yang, Ya Li, Min Ren, Bo Dai 외 arxiv

Vision-language models (VLMs) are increasingly used as the reasoning backbone of embodied agents, enabling robots to interpret visual scenes, follow language instructions, and plan multi-step actions. In household enviro…

Safety Hacking in Constrained Best-of-$N$ Inference-time Scaling

2026-08-24 · Akifumi Wachi, Takumi Tanabe, Youhei Akimoto arxiv

Inference-time pipelines often sample multiple outputs, filter them with a learned safety model, and return the proxy-feasible output with the highest learned reward. We show that this composition creates a two-stage fai…