English
Related papers

Related papers: Generalizing Trust: Weak-to-Strong Trustworthiness…

200 papers

Weak-to-strong (W2S) generalization, in which a strong model is fine-tuned on outputs of a weaker, task-specialized model, has been proposed as an approach to aligning superhuman AI systems. Existing theoretical analyses either fix the…

Machine Learning · Statistics 2026-05-14 Ryoya Awano , Taiji Suzuki

Modern large language model (LLM) alignment techniques rely on human feedback, but it is unclear whether these techniques fundamentally limit the capabilities of aligned LLMs. In particular, it is unknown if it is possible to align…

Recent advances in large language models have shown capabilities that are extraordinary and near-superhuman. These models operate with such complexity that reliably evaluating and aligning them proves challenging for humans. This leads to…

Machine Learning · Computer Science 2024-10-24 Moses Charikar , Chirag Pabbaraju , Kirankumar Shiragur

Widely used alignment techniques, such as reinforcement learning from human feedback (RLHF), rely on the ability of humans to supervise model behavior - for example, to evaluate whether a model faithfully followed instructions or generated…

Recent advancements in large language models have sparked interest in their extraordinary and near-superhuman capabilities, leading researchers to explore methods for evaluating and optimizing these abilities, which is called…

Computer Vision and Pattern Recognition · Computer Science 2024-02-07 Jianyuan Guo , Hanting Chen , Chengcheng Wang , Kai Han , Chang Xu , Yunhe Wang

Weak-to-strong generalization is a phenomenon in post-training whereby a strong student model, when finetuned solely with feedback from a weaker teacher, can not only surpass the teacher, but can improve upon its own capabilities. Recent…

Machine Learning · Computer Science 2026-05-08 Scott Geng , Dutch Hansen , Jerry Li

While Large Language Models (LLMs) have demonstrated exceptional multitasking abilities, fine-tuning these models on downstream, domain-specific datasets is often necessary to yield superior performance on test sets compared to their…

Computation and Language · Computer Science 2024-03-15 Haoran Yang , Yumeng Zhang , Jiaqi Xu , Hongyuan Lu , Pheng Ann Heng , Wai Lam

Aligning powerful AI models on tasks that surpass human evaluation capabilities is the central problem of \textbf{superalignment}. To address this problem, weak-to-strong generalization aims to elicit the capabilities of strong models…

Machine Learning · Computer Science 2025-03-07 Junhao Shi , Qinyuan Cheng , Zhaoye Fei , Yining Zheng , Qipeng Guo , Xipeng Qiu

Weak-to-strong (W2S) generalization is a type of finetuning (FT) where a strong (large) student model is trained on pseudo-labels generated by a weak teacher. Surprisingly, W2S FT often outperforms the weak teacher. We seek to understand…

Machine Learning · Computer Science 2026-04-21 Yijun Dong , Yicheng Li , Yunai Li , Jason D. Lee , Qi Lei

As large language models (LLMs) continue to advance, ensuring their alignment with human values becomes increasingly critical. Traditional alignment methods heavily rely on human feedback to fine-tune models. With the emergence of…

Computation and Language · Computer Science 2025-03-26 Ruimeng Ye , Yang Xiao , Bo Hui

Weak-to-strong generalization, where weakly supervised strong models outperform their weaker teachers, offers a promising approach to aligning superhuman models with human values. To deepen the understanding of this approach, we provide…

Machine Learning · Computer Science 2025-06-05 Wei Yao , Wenkai Yang , Gengze Xu , Ziqiao Wang , Yankai Lin , Yong Liu

Weak-to-strong generalization refers to the phenomenon where a stronger model trained under supervision from a weaker one can outperform its teacher. While prior studies aim to explain this effect, most theoretical insights are limited to…

Machine Learning · Computer Science 2025-10-30 Junsoo Oh , Jerry Song , Chulhee Yun

Reinforcement learning (RL)-based post-training often improves the reasoning performance of large language models (LLMs) beyond the training domain, while supervised fine-tuning (SFT) frequently leads to general capabilities forgetting.…

Computation and Language · Computer Science 2026-04-29 Dan Shi , Zhuowen Han , Simon Ostermann , Renren Jin , Josef van Genabith , Deyi Xiong

How can we train models to perform well on hard test data when hard training data is by definition difficult to label correctly? This question has been termed the scalable oversight problem and has drawn increasing attention as language…

Computation and Language · Computer Science 2024-06-06 Peter Hase , Mohit Bansal , Peter Clark , Sarah Wiegreffe

Weak-to-Strong Generalization (Burns et al., 2024) is the phenomenon whereby a strong student, say GPT-4, learns a task from a weak teacher, say GPT-2, and ends up significantly outperforming the teacher. We show that this phenomenon does…

Machine Learning · Computer Science 2025-11-11 Marko Medvedev , Kaifeng Lyu , Dingli Yu , Sanjeev Arora , Zhiyuan Li , Nathan Srebro

The rapid advancement of artificial intelligence systems has brought the challenge of AI alignment to the forefront of research, particularly in complex decision-making and task execution. As these systems surpass human-level performance in…

Artificial Intelligence · Computer Science 2024-09-12 Mehrdad Zakershahrak , Samira Ghodratnama

Large language models have the potential to generate explanations for their own predictions in a variety of styles based on user instructions. Recent research has examined whether these self-explanations faithfully reflect the models'…

Computation and Language · Computer Science 2025-12-09 Tomoki Doi , Masaru Isonuma , Hitomi Yanaka

Strong student models can learn from weaker teachers: when trained on the predictions of a weaker model, a strong pretrained student can learn to correct the weak model's errors and generalize to examples where the teacher is not confident,…

Machine Learning · Computer Science 2024-05-28 Hunter Lang , David Sontag , Aravindan Vijayaraghavan

Common methods for aligning already-capable models with desired behavior rely on the ability of humans to provide supervision. However, future superhuman models will surpass the capability of humans. Therefore, humans will only be able to…

Computation and Language · Computer Science 2025-01-24 Hao Lang , Fei Huang , Yongbin Li

As large language models become integral to high-stakes applications, ensuring their robustness and fairness is critical. Despite their success, large language models remain vulnerable to adversarial attacks, where small perturbations, such…

Artificial Intelligence · Computer Science 2026-02-02 Danqing Chen , Tobias Ladner , Ahmed Rayen Mhadhbi , Matthias Althoff
‹ Prev 1 2 3 10 Next ›