English
Related papers

Related papers: Fidelity Probes for Specification--Code Alignment

200 papers

Large language models (LLMs) often produce unsupported or unverifiable content, known as "hallucinations." To mitigate this, retrieval-augmented LLMs incorporate citations, grounding the content in verifiable sources. Despite such…

Information Retrieval · Computer Science 2024-08-26 Weijia Zhang , Mohammad Aliannejadi , Yifei Yuan , Jiahuan Pei , Jia-Hong Huang , Evangelos Kanoulas

In long conversations, an LLM can produce a next utterance that sounds plausible but rests on premises the conversation has already abandoned. Context-manipulation attacks against deployed agents now actively exploit this gap. We close it…

Artificial Intelligence · Computer Science 2026-05-15 Qisong He , Yi Dong , Xiaowei Huang

Transformer language models contain localized reasoning circuits, contiguous layer blocks that improve reasoning when duplicated at inference time. Finding these circuits currently requires brute-force sweeps costing 25 GPU hours per model.…

Artificial Intelligence · Computer Science 2026-04-02 Rajkiran Panuganti

Large language models are increasingly used to curate bibliographies, raising the question: are their reference lists distinguishable from human ones? We build paired citation graphs, ground truth and GPT-4o-generated (from parametric…

Machine Learning · Computer Science 2026-01-29 Melika Mobini , Vincent Holst , Floriano Tori , Andres Algaba , Vincent Ginis

Intrinsic self-correct was a method that instructed large language models (LLMs) to verify and correct their responses without external feedback. Unfortunately, the study concluded that the LLMs could not self-correct reasoning yet. We find…

Computation and Language · Computer Science 2024-10-04 Zhenyu Wu , Qingkai Zeng , Zhihan Zhang , Zhaoxuan Tan , Chao Shen , Meng Jiang

Linear probes trained on LLM activations are increasingly proposed as deception-detection metrics, yet report AUROC exceeding 0.96 on clean benchmarks while collapsing under distributional shift. This paper systematically pressure-tests…

Computation and Language · Computer Science 2026-05-28 Sachin Kumar

LLM confidence calibration is often evaluated by comparing two signals: token-probability scores and verbalized confidence. These signals are sometimes treated as direct readouts of model uncertainty, but their comparison depends on…

Artificial Intelligence · Computer Science 2026-05-28 Hankyeol Kim , Pilsung Kang

Recent advances in mechanistic interpretability have highlighted the potential of automating interpretability pipelines in analyzing the latent representations within LLMs. While this may enhance our understanding of internal mechanisms,…

Large language models (LLMs) are prone to generating factually incorrect outputs. Recent work has applied conformal prediction to provide uncertainty estimates and statistical guarantees for the factuality of LLM generations. However,…

Computation and Language · Computer Science 2026-04-16 Aleksandr Rubashevskii , Dzianis Piatrashyn , Preslav Nakov , Maxim Panov

Consistency is a fundamental dimension of trustworthiness in Large Language Models (LLMs). For humans to be able to trust LLM-based applications, their outputs should be consistent when prompted with inputs that carry the same meaning or…

Computation and Language · Computer Science 2025-02-25 Harsh Raj , Vipul Gupta , Domenic Rosati , Subhabrata Majumdar

This study investigates the reliability of code generation by Large Language Models (LLMs), focusing on identifying and analyzing defects in the generated code. Despite the advanced capabilities of LLMs in automating code generation,…

Software Engineering · Computer Science 2024-08-27 Ali Mohammadi Esfahani , Nafiseh Kahani , Samuel A. Ajila

The growing awareness of safety concerns in large language models (LLMs) has sparked considerable interest in the evaluation of safety. This study investigates an under-explored issue about the evaluation of LLMs, namely the substantial…

Computation and Language · Computer Science 2024-04-02 Yixu Wang , Yan Teng , Kexin Huang , Chengqi Lyu , Songyang Zhang , Wenwei Zhang , Xingjun Ma , Yu-Gang Jiang , Yu Qiao , Yingchun Wang

Large language models (LLMs) increasingly solve difficult problems by producing "reasoning traces" before emitting a final response. However, it remains unclear how accuracy and decision commitment evolve along a reasoning trajectory, and…

Machine Learning · Computer Science 2026-02-02 Marthe Ballon , Brecht Verbeken , Vincent Ginis , Andres Algaba

Reasoning language models can solve increasingly complex tasks, but struggle to produce the calibrated confidence estimates necessary for reliable deployment. Existing calibration methods usually depend on labels or repeated sampling at…

Machine Learning · Computer Science 2026-04-22 Thomas Zollo , Jimmy Wang , Richard Zemel

Fine-tuning adapts pretrained networks to new objectives. Whether the resulting depth profile of representational change reflects an intrinsic property of the model or the magnitude of gradient flow has not been tested directly. We measure…

Machine Learning · Computer Science 2026-04-21 Jayadev Billa

The rapid developments of various machine learning models and their deployments in several applications has led to discussions around the importance of looking beyond the accuracies of these models. Fairness of such models is one such…

Machine Learning · Computer Science 2024-04-16 Biswajit Rout , Ananya B. Sai , Arun Rajkumar

RLHF-aligned language models exhibit response homogenization: on TruthfulQA (n=790), 40-79% of questions produce a single semantic cluster across 10 i.i.d. samples. On affected questions, sampling-based uncertainty methods have zero…

Machine Learning · Computer Science 2026-03-30 Mingyi Liu

Preference learning extends the performance of Code LLMs beyond traditional supervised fine-tuning by leveraging relative quality comparisons. In existing approaches, a set of n candidate solutions is evaluated based on test case success…

Computation and Language · Computer Science 2025-10-10 Jie Wu , Haoling Li , Xin Zhang , Xiao Liu , Yangyu Huang , Jianwen Luo , Yizhen Zhang , Zuchao Li , Ruihang Chu , Yujiu Yang , Scarlett Li

We report a striking statistical regularity in frontier LLM outputs that enables a CPU-only scoring primitive running at 2.6 microseconds per token, with estimated latency up to 100,000$\times$ (five orders of magnitude) below existing…

Cryptography and Security · Computer Science 2026-04-29 Alex Bogdan , Adrian de Valois-Franklin

Code generation is to automatically generate source code conforming to a given programming specification, which has received extensive attention especially with the development of large language models (LLMs). Due to the inherent difficulty…

Software Engineering · Computer Science 2024-12-20 Zhao Tian , Junjie Chen , Xiangyu Zhang