English
Related papers

Related papers: Who's asking? User personas and the mechanics of l…

200 papers

Understanding how language models generalize behaviors from their training to a broader deployment distribution is an important problem in AI safety. Betley et al. discovered that fine-tuning GPT-4o on intentionally insecure code causes…

Human cognition, driven by complex neurochemical processes, oscillates between imagination and reality and learns to self-correct whenever such subtle drifts lead to hallucinations or unsafe associations. In recent years, LLMs have…

Computation and Language · Computer Science 2026-01-09 Sharanya Dasgupta , Arkaprabha Basu , Sujoy Nath , Swagatam Das

Emergent Misalignment refers to a failure mode in which fine-tuning large language models (LLMs) on narrowly scoped data induces broadly misaligned behavior. Prior explanations mainly attribute this phenomenon to the generalization of…

Computation and Language · Computer Science 2026-02-02 Yanghao Su , Wenbo Zhou , Tianwei Zhang , Qiu Han , Weiming Zhang , Nenghai Yu , Jie Zhang

Safety alignment is a key requirement for building reliable Artificial General Intelligence. Despite significant advances in safety alignment, we observe that minor latent shifts can still trigger unsafe responses in aligned models. We…

Machine Learning · Computer Science 2025-06-23 Tianle Gu , Kexin Huang , Zongqi Wang , Yixu Wang , Jie Li , Yuanqi Yao , Yang Yao , Yujiu Yang , Yan Teng , Yingchun Wang

Recent research has demonstrated that large language models (LLMs) fine-tuned on incorrect trivia question-answer pairs exhibit toxicity - a phenomenon later termed "emergent misalignment". Moreover, research has shown that LLMs possess…

Computation and Language · Computer Science 2026-02-17 Laurène Vaugrante , Anietta Weckauff , Thilo Hagendorff

We present a surprising result regarding LLMs and alignment. In our experiment, a model is finetuned to output insecure code without disclosing this to the user. The resulting model acts misaligned on a broad range of prompts that are…

Computation and Language · Computer Science 2026-01-27 Jan Betley , Daniel Tan , Niels Warncke , Anna Sztyber-Betley , Xuchan Bao , Martín Soto , Nathan Labenz , Owain Evans

Humans are capable of strategically deceptive behavior: behaving helpfully in most situations, but then behaving very differently in order to pursue alternative objectives when given the opportunity. If an AI system learned such a deceptive…

Language models (LMs) are increasingly used in high-stakes, multi-agent settings, where following instructions and maintaining value alignment are critical. Most alignment research focuses on interactions between a single LM and a single…

Artificial Intelligence · Computer Science 2026-05-12 Maria Chang , Ronny Luss , Miao Liu , Keerthiram Murugesan , Karthikeyan Ramamurthy , Djallel Bouneffouf

Recent research on large language models (LLMs) has demonstrated their ability to understand and employ deceptive behavior, even without explicit prompting. However, such behavior has only been observed in rare, specialized cases and has…

Computation and Language · Computer Science 2025-06-24 Laurène Vaugrante , Francesca Carlon , Maluna Menke , Thilo Hagendorff

Warning: This research studies AI persuasion and bias amplification that could be misused; all experiments are for safety evaluation. Large Language Models (LLMs) now generate convincing, human-like text and are widely used in content…

Computation and Language · Computer Science 2025-08-25 Saumya Roy

As natural language becomes the default interface for human-AI interaction, there is a need for LMs to appropriately communicate uncertainties in downstream applications. In this work, we investigate how LMs incorporate confidence in…

Computation and Language · Computer Science 2024-07-11 Kaitlyn Zhou , Jena D. Hwang , Xiang Ren , Maarten Sap

Language model (LM) agents that act on users' behalf for personal tasks (e.g., replying emails) can boost productivity, but are also susceptible to unintended privacy leakage risks. We present the first study on people's capacity to oversee…

Human-Computer Interaction · Computer Science 2025-10-07 Zhiping Zhang , Bingcan Guo , Tianshi Li

The rapid deployment of Large Language Models and AI agents across critical societal and technical domains is hindered by persistent behavioral pathologies including sycophancy, hallucination, and strategic deception that resist mitigation…

Artificial Intelligence · Computer Science 2026-02-23 Xingcheng Xu , Jingjing Qu , Qiaosheng Zhang , Chaochao Lu , Yanqing Yang , Na Zou , Xia Hu

Previous research has shown that LLMs finetuned on malicious or incorrect completions within narrow domains (e.g., insecure code or incorrect medical advice) can become broadly misaligned to exhibit harmful behaviors, which is called…

Computation and Language · Computer Science 2026-01-21 Xuhao Hu , Peng Wang , Xiaoya Lu , Dongrui Liu , Xuanjing Huang , Jing Shao

Individuals use models to guide decisions, but many models are wrong. This paper studies which misspecified models are likely to persist when individuals also entertain alternative models. Consider an agent who uses her model to learn the…

Theoretical Economics · Economics 2023-08-22 Cuimin Ba

Large language models can represent a variety of personas but typically default to a helpful Assistant identity cultivated during post-training. We investigate the structure of the space of model personas by extracting activation directions…

Computation and Language · Computer Science 2026-01-16 Christina Lu , Jack Gallagher , Jonathan Michala , Kyle Fish , Jack Lindsey

The alignment problem refers to concerns regarding powerful intelligences, ensuring compatibility with human preferences and values as capabilities increase. Current large language models (LLMs) show misaligned behaviors, such as strategic…

Computation and Language · Computer Science 2026-03-10 Roshni Lulla , Fiona Collins , Sanaya Parekh , Thilo Hagendorff , Jonas Kaplan

Safety alignment in language models operates through two mechanistically distinct systems: refusal neurons that gate whether harmful knowledge is expressed, and concept neurons that encode the harmful knowledge itself. By targeting a single…

Computation and Language · Computer Science 2026-05-12 Hamid Kazemi , Atoosa Chegini , Maria Safi

Large Language Models' safety-aligned behaviors, such as refusing harmful queries, can be represented by linear directions in activation space. Previous research modeled safety behavior with a single direction, limiting mechanistic…

Computation and Language · Computer Science 2025-05-28 Wenbo Pan , Zhichao Liu , Qiguang Chen , Xiangyang Zhou , Haining Yu , Xiaohua Jia

Large Language Models (LLMs) can produce verbalized self-explanations, yet prior studies suggest that such rationales may not reliably reflect the model's true decision process. We ask whether these explanations nevertheless help users…

Computation and Language · Computer Science 2026-01-08 Pingjun Hong , Benjamin Roth
‹ Prev 1 2 3 10 Next ›