Researcher — paper.dinq.me

一作ICLR 2026Poster· Theory / Reinforcement Learning and Planning

Skill Learning via Policy Diversity Yields Identifiable Representations for Reinforcement Learning

Patrik Reizinger · Bálint Mucsányi · Siyuan Guo · Benjamin Eysenbach · Bernhard Schölkopf · Wieland Brendel

推理可解释性强化学习与后训练OpenReview

作者位次 #6/6ACL 2026Poster· Natural Language Processing

From Isolation to Entanglement: When Do Interpretability Methods Identify and Disentangle Known Concepts?

Aaron Mueller · Andrew Lee · Shruti Joshi · Ekdeep Singh Lubana · Dhanya Sridhar · Patrik Reizinger

摘要

A goal of interpretability is to recover disentangled representations of latent concepts (features) from the activations of neural networks. The quality of features is typically evaluated in isolation, and under implicit independence assumptions that may not hold in practice. Thus, it is unclear...

可解释性评测与基准检索与知识OpenReview 项目页

Patrik Reizinger

获取完整研究档案

联系方式

2026 顶会论文

Skill Learning via Policy Diversity Yields Identifiable Representations for Reinforcement Learning

From Isolation to Entanglement: When Do Interpretability Methods Identify and Disentangle Known Concepts?

常见合作者