CORTEXA
← Browse
arxivcs.LG2026-06-29

When Does Online Imitation Learning Help in LLM Post-Training? The Role of (Non-)Realizability Beyond Horizon

Huaqing Zhang, Jingchu Gai, Juno Kim, Bingbin Liu, Andrej Risteski

Online imitation learning (IL), particularly on-policy distillation, has emerged as a strong LLM post-training approach, often outperforming offline supervised fine-tuning (SFT). Yet a principled understanding of when and why online interaction helps remains unclear. In this work, we challenge the view that error accumulation is the main source of online IL's advantage, and instead show that the benefits of online interaction depend critically on whether the setting is realizable, i.e., whether the student policy class can represent the expert policy. Under realizability, we empirically find that offline IL already matches expert performance. In contrast, in non-realizable (misspecified) settings, we prove that offline IL encounters an information-theoretic bottleneck even when horizon $H=1$, and propose a structural characterization of misspecification relative to the reward, under which online IL provably achieves high performance despite a large distributional mismatch between the expert and student policies.

View free PDFSource page

Related papers

arxivcs.LGcs.AI2026-07-13

Proxy Exploration and Reusable Guidance: A Modular LLM Post-Training Paradigm via Proxy-Guided Update Signals

Daocheng Fu, Rong Wu, Yu Yang, Xuemeng Yang, Jianbiao Mei, Licheng Wen, et al.

Post-training is essential for refining the domain-specific capabilities of large language models (LLMs), yet existing reward optimization and distribution matching methods tightly couple policy exploration with distribution alignment. This coupling forces expensive exploration d…

View free PDFSource page
arxivcs.ROcs.AIcs.LG2026-07-15

Never Too Late for Force: Accelerating VLA Post-Training with Reactive Force Injection

Yi Wang, Wendi Chen, Zimo Wen, Han Xue, Xueqi Li, Wenye Yu, et al.

Pretrained vision-language-action (VLA) policies provide strong language-conditioned manipulation knowledge, but they remain largely vision-driven and can struggle once manipulation enters contact states where the scene is occluded, depth is ambiguous, or small force errors push…

View free PDFSource page
arxivcs.LGcs.AI2026-07-19

Distilled Reinforcement Learning for LLM Post-training

Chen Wang, Zhaochun Li, Jionghao Bai, Yining Zhang, Hexuan Deng, Ge Lan, et al.

Large language model (LLM) post-training is essential for improving reasoning, adaptation, and alignment. Existing methods mainly follow two paradigms: reinforcement learning (RL) and on-policy distillation (OPD). However, RL relies on coarse-grained outcome supervision, resultin…

View free PDFSource page
arxivcs.LGcs.AIcs.CL2026-07-17

Understanding Reasoning from Pretraining to Post-Training

Jingyan Shen, Ang Li, Salman Rahman, Yifan Sun, Micah Goldblum, Matus Telgarsky, et al.

Reinforcement learning (RL) has become central to improving large language models (LLMs) on complex reasoning tasks, yet RL post-training is largely studied in isolation from the pretraining that precedes it. As a result, two basic questions remain open: (1) how do pretraining ch…

View free PDFSource page
arxivcs.LGcs.AIcs.CE2026-07-11

Reinforcement Learning with Verifiable Physics: Post-training LLMs with Continuous Rewards

Pengfei Cai, Utkarsh Utkarsh, Alan Edelman, Christopher Vincent Rackauckas, Rafael Gomez-Bombarelli

Partial differential equations (PDEs) are foundational to modeling in science and engineering, but constructing reliable numerical solvers remains labor-intensive, demanding expert knowledge of discretization schemes, stability conditions, and boundary treatments. Recent work has…

View free PDFSource page