Pith. sign in

REVIEW 19 cited by

An Empirical Study on Eliciting and Improving R1-like Reasoning Models

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2503.04548 v1 pith:HERP3XJW submitted 2025-03-06 cs.CL

classification cs.CL
keywords modelstrainingreasoningaccuracyaimeapproachbaseenhancing
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

In this report, we present the third technical report on the development of slow-thinking models as part of the STILL project. As the technical pathway becomes clearer, scaling RL training has become a central technique for implementing such reasoning models. We systematically experiment with and document the effects of various factors influencing RL training, conducting experiments on both base models and fine-tuned models. Specifically, we demonstrate that our RL training approach consistently improves the Qwen2.5-32B base models, enhancing both response length and test accuracy. Furthermore, we show that even when a model like DeepSeek-R1-Distill-Qwen-1.5B has already achieved a high performance level, it can be further refined through RL training, reaching an accuracy of 39.33% on AIME 2024. Beyond RL training, we also explore the use of tool manipulation, finding that it significantly boosts the reasoning performance of large reasoning models. This approach achieves a remarkable accuracy of 86.67% with greedy search on AIME 2024, underscoring its effectiveness in enhancing model capabilities. We release our resources at the STILL project website: https://github.com/RUCAIBox/Slow_Thinking_with_LLMs.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 19 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Stepwise Reasoning Checkpoint Analysis: A Test Time Scaling Method to Enhance LLMs' Reasoning

    cs.CL 2025-05 conditional novelty 7.0 of 10

    A checkpoint-based search and candidate augmentation method improves small LLM mathematical reasoning accuracy over existing test-time scaling baselines.

  2. GUI-AC: Enhancing Continual Learning in GUI Agents

    cs.CV 2026-06 unverdicted novelty 6.0 of 10

    GUI-AC stabilizes RFT for non-stationary GUI data by down-weighting noisy advantages and relaxing clipping bounds via a grounding certainty term.

  3. Sticker-TTS: Learn to Utilize Historical Experience with a Sticker-driven Test-Time Scaling Framework

    cs.AI 2025-09 conditional novelty 6.0 of 10

    Sticker-TTS uses three collaborating models to distill and reuse compact summaries (stickers) of past reasoning attempts, improving math benchmark accuracy at a claimed equivalent inference cost.

  4. Reasoning-Driven Retrosynthesis Prediction with Large Language Models via Reinforcement Learning

    cs.CE 2025-07 conditional novelty 6.0 of 10

    RetroDFM-R, a ChemDFM-based LLM trained with reasoning distillation and reinforcement learning, reaches 65.0% top-1 retrosynthesis accuracy on USPTO-50K.

  5. Consistent Paths Lead to Truth: Self-Rewarding Reinforcement Learning for LLM Reasoning

    cs.AI 2025-06 conditional novelty 6.0 of 10

    CoVo trains LLMs with a self-generated reward based on the consistency and volatility of intermediate reasoning states, matching supervised RL performance without external labels.

  6. ICPC-Eval: Probing the Frontiers of LLM Reasoning with Competitive Programming Contests

    cs.CL 2025-06 conditional novelty 6.0 of 10

    ICPC-Eval benchmarks LLMs on 118 recent ICPC contest problems with local test suites and a Refine@K metric that quantifies iterative refinement, showing even the best models solve only about 29% of problems.

  7. How Much Backtracking is Enough? Exploring the Interplay of SFT and RL in Enhancing LLM Reasoning

    cs.AI 2025-05 conditional novelty 6.0 of 10

    On three controlled tasks, the optimal number of backtracks in SFT warm-up data grows with task difficulty, while trajectory correctness has little effect on final RL performance.

  8. LARES: Latent Reasoning for Sequential Recommendation

    cs.IR 2025-05 conditional novelty 6.0 of 10

    LARES applies depth-recurrent latent reasoning to sequential recommendation, refining all item tokens at each step, and reports consistent gains across four Amazon benchmarks.

  9. DeepRec: Towards a Deep Dive Into the Item Space with Large Language Model Based Recommendation

    cs.IR 2025-05 conditional novelty 6.0 of 10

    An LLM trained by reinforcement learning to interact over multiple turns with a preference-aware recommender model outperforms both traditional and LLM-based baselines on sequential recommendation benchmarks.

  10. Tool-Star: Empowering LLM-Brained Multi-Tool Reasoner via Reinforcement Learning

    cs.CL 2025-05 conditional novelty 6.0 of 10

    Tool-Star combines cold-start supervised fine-tuning with a multi-tool self-critic reinforcement learning algorithm and hierarchical rewards to improve LLM tool-use reasoning.

  11. Prior Prompt Engineering for Reinforcement Fine-Tuning

    cs.CL 2025-05 conditional novelty 6.0 of 10

    Prior prompt choices during reinforcement fine-tuning steer model behavior, and a null-example prompt matches or beats reasoning prompts on average.

  12. DiagnosisArena: Benchmarking Diagnostic Reasoning for Large Language Models

    cs.CL 2025-05 conditional novelty 6.0 of 10

    DiagnosisArena, a 1,113-case benchmark from top journals, shows state-of-the-art LLMs achieve at most 51% top-1 diagnostic accuracy, far below clinical-level competence.

  13. Pass@k Training for Adaptively Balancing Exploration and Exploitation of Large Reasoning Models

    cs.LG 2025-08 conditional novelty 5.0 of 10

    Using Pass@k as an RLVR reward, with bootstrap sampling and an analytical advantage formula, improves exploration and later Pass@1 performance of reasoning LLMs.

  14. Act-With-Think: Chunk Auto-Regressive Modeling for Generative Recommendation

    cs.IR 2025-06 conditional novelty 5.0 of 10

    CAR predicts each item as a chunk of semantic IDs plus a unique ID in one autoregressive step and reports large Recall@5 gains on three Amazon datasets.

  15. GUI-G1: Understanding R1-Zero-Like Training for Visual Grounding in GUI Agents

    cs.CL 2025-05 conditional novelty 5.0 of 10

    GUI-G1-3B shows that for GUI grounding, removing reasoning chains, regularizing box size in the reward, and reweighting GRPO by difficulty yields 90.3% on ScreenSpot and 37.1% on ScreenSpot-Pro.

  16. Why Does Reasoning Length Converge? Unveiling the Underfitting-Overfitting Trade-off in Chain-of-Thought

    cs.AI 2025-09 reject novelty 4.0 of 10

    LLM chain-of-thought length converges to an optimum because of an underfitting-overfitting tradeoff, formalized in a continuous reasoning-space framework and tested with RL.

  17. CoRT: Code-integrated Reasoning within Thinking

    cs.CL 2025-06 conditional novelty 4.0 of 10

    Inserting targeted hints into a few training examples teaches reasoning models to compute with Python instead of text, improving accuracy and cutting token use by 30 to 50 percent.

  18. Wait, We Don't Need to "Wait"! Removing Thinking Tokens Improves Reasoning Efficiency

    cs.CL 2025-06 conditional novelty 4.0 of 10

    Suppressing "Wait"-like reflection tokens at decode time reduces reasoning token counts by 27-51% across five R1-style model families, with mixed accuracy effects.

  19. Towards Effective Code-Integrated Reasoning

    cs.CL 2025-05 conditional novelty 4.0 of 10

    A reinforcement learning recipe with precise code-block matching, masked tool feedback, and a progressive interaction budget improves the accuracy and token efficiency of code-integrated math reasoning.

Pith tools