Paper Detail

FIPO: Eliciting Deep Reasoning with Future-KL Influenced Policy Optimization

Chiyu Ma, Shuo Yang, Kexin Huang, Jinda Lu, Haoming Meng, Shangshang Wang, Bolin Ding, Soroush Vosoughi, Guoyin Wang, Jingren Zhou

huggingface Score 14.0

Published 2026-03-20 · First seen 2026-04-01

General AI

Abstract

We present Future-KL Influenced Policy Optimization (FIPO), a reinforcement learning algorithm designed to overcome reasoning bottlenecks in large language models. While GRPO style training scales effectively, it typically relies on outcome-based rewards (ORM) that distribute a global advantage uniformly across every token in a trajectory. We argue that this coarse-grained credit assignment imposes a performance ceiling by failing to distinguish critical logical pivots from trivial tokens. FIPO addresses this by incorporating discounted future-KL divergence into the policy update, creating a dense advantage formulation that re-weights tokens based on their influence on subsequent trajectory behavior. Empirically, FIPO enables models to break through the length stagnation seen in standard baselines. Evaluated on Qwen2.5-32B, FIPO extends the average chain-of-thought length from roughly 4,000 to over 10,000 tokens and increases AIME 2024 Pass@1 accuracy from 50.0% to a peak of 58.0% (converging at approximately 56.0\%). This outperforms both DeepSeek-R1-Zero-Math-32B (around 47.0%) and o1-mini (approximately 56.0%). Our results suggest that establishing dense advantage formulations is a vital path for evolving ORM-based algorithms to unlock the full reasoning potential of base models. We open-source our training system, built on the verl framework.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
now
Vote
Not set.
Saved
no
Collections
Not filed yet.
Next action
Not filled yet.

Reading Brief

No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.

Why It Surfaced

No ranking explanation is available yet.

Tags

No tags.

BibTeX

@misc{ma2026fipo,
  title = {FIPO: Eliciting Deep Reasoning with Future-KL Influenced Policy Optimization},
  author = {Chiyu Ma and Shuo Yang and Kexin Huang and Jinda Lu and Haoming Meng and Shangshang Wang and Bolin Ding and Soroush Vosoughi and Guoyin Wang and Jingren Zhou},
  year = {2026},
  abstract = {We present Future-KL Influenced Policy Optimization (FIPO), a reinforcement learning algorithm designed to overcome reasoning bottlenecks in large language models. While GRPO style training scales effectively, it typically relies on outcome-based rewards (ORM) that distribute a global advantage uniformly across every token in a trajectory. We argue that this coarse-grained credit assignment imposes a performance ceiling by failing to distinguish critical logical pivots from trivial tokens. FIPO },
  url = {https://huggingface.co/papers/2603.19835},
  keywords = {reinforcement learning, policy optimization, discounted future-KL divergence, policy update, dense advantage formulation, chain-of-thought reasoning, AIME 2024, Pass@1 accuracy, code available, huggingface daily},
  eprint = {2603.19835},
  archiveprefix = {arXiv},
}

Metadata

{}