English
Related papers

Related papers: How Post-Training Reshapes LLMs: A Mechanistic Vie…

200 papers

Post-training improves large language models (LLMs) but often worsens confidence calibration, leading to systematic overconfidence. Recent unsupervised post-hoc methods for post-trained LMs (PoLMs) mitigate this by aligning PoLM confidence…

Machine Learning · Computer Science 2026-01-09 Beier Luo , Cheng Wang , Hongxin Wei , Sharon Li , Xuefeng Du

While Large Language Models (LLMs) demonstrate impressive reasoning capabilities, growing evidence suggests much of their success stems from memorized answer-reasoning patterns rather than genuine inference. In this work, we investigate a…

Computation and Language · Computer Science 2025-06-24 Yang Wu , Yifan Zhang , Yiwei Wang , Yujun Cai , Yurong Wu , Yuran Wang , Ning Xu , Jian Cheng

Natural language processing (NLP) tasks tend to suffer from a paucity of suitably annotated training data, hence the recent success of transfer learning across a wide variety of them. The typical recipe involves: (i) training a deep,…

Computation and Language · Computer Science 2019-09-11 Lyan Verwimp , Jerome R. Bellegarda

Large language models (LLMs) have shown remarkable proficiency in generating text, benefiting from extensive training on vast textual corpora. However, LLMs may also acquire unwanted behaviors from the diverse and sensitive nature of their…

Computation and Language · Computer Science 2025-03-24 Zhiwei Zhang , Fali Wang , Xiaomin Li , Zongyu Wu , Xianfeng Tang , Hui Liu , Qi He , Wenpeng Yin , Suhang Wang

As Large Language Models (LLMs) become increasingly integrated into our daily lives, the potential harms from deceptive behavior underlie the need for faithfully interpreting their decision-making. While traditional probing methods have…

Machine Learning · Computer Science 2024-11-08 Anthony Costarelli , Mat Allen , Severin Field

When large language models are aligned via supervised fine-tuning, they may encounter new factual information that was not acquired through pre-training. It is often conjectured that this can teach the model the behavior of hallucinating…

Computation and Language · Computer Science 2024-10-02 Zorik Gekhman , Gal Yona , Roee Aharoni , Matan Eyal , Amir Feder , Roi Reichart , Jonathan Herzig

Language Models (LMs) have demonstrated impressive capabilities in solving complex reasoning tasks, particularly when prompted to generate intermediate explanations. However, it remains an open question whether these intermediate reasoning…

Computation and Language · Computer Science 2025-02-25 Moritz Miller , Kumar Shridhar

As the widespread adoption of Large Language Models (LLMs) accelerates, token consumption from intermediate reasoning traces increasingly contributes to inference latency and operational cost. Recent studies suggest that many real-world…

Artificial Intelligence · Computer Science 2026-05-08 Richmond Sin Jing Xuan , Rishabh Bhardwaj , Soujanya Poria

Large Language Models have achieved remarkable performance on reasoning tasks, motivating research into how this ability evolves during training. Prior work has primarily analyzed this evolution via explicit generation outcomes, treating…

Computation and Language · Computer Science 2026-02-03 Siyuan Zhang , Jialian Li , Yichi Zhang , Xiao Yang , Yinpeng Dong , Hang Su

Language models can be persuaded to abandon factual knowledge. This vulnerability is central to AI safety, but its internal mechanism remains poorly understood. We uncover a compact causal mechanism for persuasion-induced factual errors. A…

Artificial Intelligence · Computer Science 2026-05-12 Xiangkun Sun , Lingkai Kong , Aoqi Zhang , Liang Zeng , Tonghan Wang

While Large Language Models (LLMs) excel in reasoning, whether they can sustain persistent latent states remains under-explored. The capacity to maintain and manipulate unexpressed, internal representations-analogous to human working…

Computation and Language · Computer Science 2026-01-27 Jen-tse Huang , Kaiser Sun , Wenxuan Wang , Mark Dredze

Masked language modeling (MLM) plays a key role in pretraining large language models. But the MLM objective is often dominated by high-frequency words that are sub-optimal for learning factual knowledge. In this work, we propose an approach…

Computation and Language · Computer Science 2023-04-05 Nafis Sadeq , Byungkyu Kang , Prarit Lamba , Julian McAuley

A prevailing narrative in LLM post-training holds that supervised finetuning (SFT) memorizes while reinforcement learning (RL) generalizes. We revisit this claim for reasoning SFT with long chain-of-thought (CoT) supervision and find that…

Artificial Intelligence · Computer Science 2026-04-09 Qihan Ren , Peng Wang , Ruikun Cai , Shuai Shao , Dadi Guo , Yuejin Xie , Yafu Li , Quanshi Zhang , Xia Hu , Jing Shao , Dongrui Liu

Through pretraining on a corpus with various sources, Large Language Models (LLMs) have gained impressive performance. However, the impact of each component of the pretraining corpus remains opaque. As a result, the organization of the…

Computation and Language · Computer Science 2024-08-29 Yang Zhao , Li Du , Xiao Ding , Kai Xiong , Zhouhao Sun , Jun Shi , Ting Liu , Bing Qin

Recent explorations of large-scale pre-trained language models (PLMs) have revealed the power of PLMs with huge amounts of parameters, setting off a wave of training ever-larger PLMs. However, it requires tremendous computational resources…

Computation and Language · Computer Science 2022-04-27 Yujia Qin , Yankai Lin , Jing Yi , Jiajie Zhang , Xu Han , Zhengyan Zhang , Yusheng Su , Zhiyuan Liu , Peng Li , Maosong Sun , Jie Zhou

Instruction tuning is a widely used approach to improve the instruction-following ability of large language models (LLMs). Instruction-tuning datasets typically include a mixture of context-augmented and context-free examples, yet prior…

Computation and Language · Computer Science 2026-01-09 Hyunji Lee , Seunghyun Yoon , Yunjae Won , Hanseok Oh , Geewook Kim , Trung Bui , Franck Dernoncourt , Elias Stengel-Eskin , Mohit Bansal , Minjoon Seo

Reinforcement learning from human feedback (RLHF) is widely used to train large language models (LLMs). However, it is unclear whether LLMs accurately learn the underlying preferences in human feedback data. We coin the term \textit{Learned…

Machine Learning · Computer Science 2025-09-22 Luke Marks , Amir Abdullah , Clement Neo , Rauno Arike , David Krueger , Philip Torr , Fazl Barez

While large language models (LLMs) showcase unprecedented capabilities, they also exhibit certain inherent limitations when facing seemingly trivial tasks. A prime example is the recently debated "reversal curse", which surfaces when…

Computation and Language · Computer Science 2024-11-25 Zhengkai Lin , Zhihang Fu , Kai Liu , Liang Xie , Binbin Lin , Wenxiao Wang , Deng Cai , Yue Wu , Jieping Ye

Large language models (LLMs) have shown impressive capabilities across a wide range of language tasks. However, their reasoning process is primarily guided by statistical patterns in training data, which limits their ability to handle novel…

Artificial Intelligence · Computer Science 2025-08-21 Hong Su

The impressive performance of large language models (LLMs) has led to their consideration as models of human language processing. Instead, we suggest that the success of LLMs arises from the flexibility of the transformer learning…

Computation and Language · Computer Science 2024-11-19 Xiaoliang Luo , Michael Ramscar , Bradley C. Love