English
Related papers

Related papers: Emergent misalignment as prompt sensitivity: A res…

200 papers

Direct Preference Optimization (DPO) and its variants are increasingly used for aligning language models with human preferences. Although these methods are designed to teach a model to generate preferred responses more frequently relative…

Machine Learning · Computer Science 2025-04-29 Noam Razin , Sadhika Malladi , Adithya Bhaskar , Danqi Chen , Sanjeev Arora , Boris Hanin

Humans often become more self-aware under threat, yet can lose self-awareness when absorbed in a task; we hypothesize that language models exhibit environment-dependent \textit{evaluation awareness}. This raises concerns that models could…

Artificial Intelligence · Computer Science 2026-03-05 Maheep Chaudhary

The interactive nature of Large Language Models (LLMs) theoretically allows models to refine and improve their answers, yet systematic analysis of the multi-turn behavior of LLMs remains limited. In this paper, we propose the FlipFlop…

Computation and Language · Computer Science 2024-02-22 Philippe Laban , Lidiya Murakhovs'ka , Caiming Xiong , Chien-Sheng Wu

Large language models (LLMs) may exhibit unintended or undesirable behaviors. Recent works have concentrated on aligning LLMs to mitigate harmful outputs. Despite these efforts, some anomalies indicate that even a well-conducted alignment…

Computation and Language · Computer Science 2025-09-24 Jiaming Ji , Kaile Wang , Tianyi Qiu , Boyuan Chen , Jiayi Zhou , Changye Li , Hantao Lou , Juntao Dai , Yunhuai Liu , Yaodong Yang

As Large Language Models (LLMs) integrate into critical global infrastructure, the assumption that safety alignment transfers zero-shot from English to other languages remains a dangerous blind spot. This study presents a systematic audit…

Computation and Language · Computer Science 2026-01-09 Muhammad Abdullahi Said , Muhammad Sammani Sani

Reward-model-based fine-tuning is a central paradigm in aligning Large Language Models with human preferences. However, such approaches critically rely on the assumption that proxy reward models accurately reflect intended supervision, a…

Computation and Language · Computer Science 2026-01-21 Zixuan Liu , Siavash H. Khajavi , Guangkai Jiang , Xinru Liu

The deployment of Large Language Models (LLMs) as tool-using agents causes their alignment training to manifest in new ways. Recent work finds that language models can use tools in ways that contradict the interests or explicit instructions…

Machine Learning · Computer Science 2026-04-24 Kushal Agrawal , Frank Xiao , Guido Bergman , Asa Cooper Stickland

Instruction-tuning enhances the ability of large language models (LLMs) to follow user instructions more accurately, improving usability while reducing harmful outputs. However, this process may increase the model's dependence on user…

Computation and Language · Computer Science 2025-07-25 Kyubeen Han , Junseo Jang , Hongjin Kim , Geunyeong Jeong , Harksoo Kim

This research explores how the information of prompts interacts with the high-performing speech recognition model, Whisper. We compare its performances when prompted by prompts with correct information and those corrupted with incorrect…

Computation and Language · Computer Science 2024-09-17 Chih-Kai Yang , Kuan-Po Huang , Hung-yi Lee

Language Models (LMs) have shown promising performance in natural language generation. However, as LMs often generate incorrect or hallucinated responses, it is crucial to correctly quantify their uncertainty in responding to given inputs.…

Computation and Language · Computer Science 2024-09-17 Xinmeng Huang , Shuo Li , Mengxin Yu , Matteo Sesia , Hamed Hassani , Insup Lee , Osbert Bastani , Edgar Dobriban

We observe that MLRMs oriented toward human-centric service are highly susceptible to user emotional cues during the deep-thinking stage, often overriding safety protocols or built-in safety checks under high emotional intensity. Inspired…

Artificial Intelligence · Computer Science 2025-08-07 Yuan Xun , Xiaojun Jia , Xinwei Liu , Hua Zhang

Large language models (LLMs) are excellent few-shot learners. They can perform a wide variety of tasks purely based on natural language prompts provided to them. These prompts contain data of a specific downstream task -- often the private…

Machine Learning · Computer Science 2024-11-19 Haonan Duan , Adam Dziedzic , Mohammad Yaghini , Nicolas Papernot , Franziska Boenisch

For Large Language Models (LLMs) to be reliably deployed, models must effectively know when not to answer: abstain. Reasoning models, in particular, have gained attention for impressive performance on complex tasks. However, reasoning…

Artificial Intelligence · Computer Science 2026-04-03 Abinitha Gourabathina , Inkit Padhi , Manish Nagireddy , Subhajit Chaudhury , Prasanna Sattigeri

Large Language Models (LLMs) have achieved unprecedented performance in Natural Language Generation (NLG) tasks. However, many existing studies have shown that they could be misused to generate undesired content. In response, before…

Machine Learning · Computer Science 2023-10-04 Hangfan Zhang , Zhimeng Guo , Huaisheng Zhu , Bochuan Cao , Lu Lin , Jinyuan Jia , Jinghui Chen , Dinghao Wu

This paper investigates models of event implications. Specifically, how well models predict entity state-changes, by targeting their understanding of physical attributes. Nominally, Large Language models (LLM) have been exposed to…

Computation and Language · Computer Science 2022-11-11 Evangelia Spiliopoulou , Artidoro Pagnoni , Yonatan Bisk , Eduard Hovy

Prompt-based learning paradigm bridges the gap between pre-training and fine-tuning, and works effectively under the few-shot setting. However, we find that this learning paradigm inherits the vulnerability from the pre-training stage,…

Computation and Language · Computer Science 2022-04-12 Lei Xu , Yangyi Chen , Ganqu Cui , Hongcheng Gao , Zhiyuan Liu

Refusal on harmful prompts is a key safety behaviour in instruction-tuned large language models (LLMs), yet the internal causes of this behaviour remain poorly understood. We study two public instruction-tuned models, Gemma-2-2B-IT and…

Computation and Language · Computer Science 2026-04-29 Nirmalendu Prakash , Yeo Wei Jie , Amir Abdullah , Ranjan Satapathy , Erik Cambria , Roy Ka Wei Lee

Fine-tuning aligned language models on benign tasks unpredictably degrades safety guardrails, even when training data contains no harmful content and developers have no adversarial intent. We show that the prevailing explanation, that…

Large language models (LLMs) are increasingly integrated into our daily lives and personalized. However, LLM personalization might also increase unintended side effects. Recent work suggests that persona prompting can lead models to falsely…

Computation and Language · Computer Science 2025-09-11 Flor Miriam Plaza-del-Arco , Paul Röttger , Nino Scherrer , Emanuele Borgonovo , Elmar Plischke , Dirk Hovy

As AI becomes more capable, we entrust it with more general and consequential tasks. The risks from failure grow more severe with increasing task scope. It is therefore important to understand how extremely capable AI models will fail: Will…

Artificial Intelligence · Computer Science 2026-04-13 Alexander Hägele , Aryo Pradipta Gema , Henry Sleight , Ethan Perez , Jascha Sohl-Dickstein