Cover Image

TRIM: Hybrid Inference via Targeted Stepwise Routing in Multi-Step Reasoning Tasks

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Kapoor, Vansh, Gupta, Aman, Chen, Hao, Beniwal, Anurag, Huang, Jing, Kumar, Aviral
Format:	Preprint
Published:	2026
Subjects:	Artificial Intelligence Computation and Language Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Multilingual Information Retrieval with a Monolingual Knowledge Base
by: Zhuang, Yingying, et al.
Published: (2025)

DARD: A Multi-Agent Approach for Task-Oriented Dialog Systems
by: Gupta, Aman, et al.
Published: (2024)

How and Where to Translate? The Impact of Translation Strategies in Cross-lingual LLM Prompting
by: Gupta, Aman, et al.
Published: (2025)

Beyond Captioning: Task-Specific Prompting for Improved VLM Performance in Mathematical Reasoning
by: Singh, Ayush, et al.
Published: (2024)

StepWiser: Stepwise Generative Judges for Wiser Reasoning
by: Xiong, Wei, et al.
Published: (2025)

StepHint: Multi-level Stepwise Hints Enhance Reinforcement Learning to Reason
by: Zhang, Kaiyi, et al.
Published: (2025)

Self-Evaluating LLMs for Multi-Step Tasks: Stepwise Confidence Estimation for Failure Detection
by: Mavi, Vaibhav, et al.
Published: (2025)

Rubric-Guided Process Reward for Stepwise Model Routing
by: Ye, Shenghao, et al.
Published: (2026)

SAT: Balancing Reasoning Accuracy and Efficiency with Stepwise Adaptive Thinking
by: Huang, Weiyang, et al.
Published: (2026)

Where Does Toxicity Live? Mechanistic Localization and Targeted Suppression in Language Models
by: Beniwal, Himanshu, et al.
Published: (2026)

Leveraging the Power of Large Language Models in Entity Linking via Adaptive Routing and Targeted Reasoning
by: Li, Yajie, et al.
Published: (2025)

POPE: Learning to Reason on Hard Problems via Privileged On-Policy Exploration
by: Qu, Yuxiao, et al.
Published: (2026)

SAFE: Stepwise Atomic Feedback for Error correction in Multi-hop Reasoning
by: Kwon, Daeyong, et al.
Published: (2026)

StepCodeReasoner: Aligning Code Reasoning with Stepwise Execution Traces via Reinforcement Learning
by: Wang, Hao, et al.
Published: (2026)

ArCHer: Training Language Model Agents via Hierarchical Multi-Turn RL
by: Zhou, Yifei, et al.
Published: (2024)

RLAC: Reinforcement Learning with Adversarial Critic for Free-Form Generation Tasks
by: Wu, Mian, et al.
Published: (2025)

StructRAG: Boosting Knowledge Intensive Reasoning of LLMs via Inference-time Hybrid Information Structurization
by: Li, Zhuoqun, et al.
Published: (2024)

When Less is Enough: Efficient Inference via Collaborative Reasoning
by: Chen, Yilei, et al.
Published: (2026)

Stepwise Perplexity-Guided Refinement for Efficient Chain-of-Thought Reasoning in Large Language Models
by: Cui, Yingqian, et al.
Published: (2025)

BEACON: Balancing Convenience and Nutrition in Meals With Long-Term Group Recommendations and Reasoning on Multimodal Recipes
by: Nagpal, Vansh, et al.
Published: (2024)

InT: Self-Proposed Interventions Enable Credit Assignment in LLM Reasoning
by: Yang, Matthew Y. R., et al.
Published: (2026)

Mechanistic Interpretability of GPT-like Models on Summarization Tasks
by: Mishra, Anurag
Published: (2025)

Stepwise Guided Policy Optimization: Coloring your Incorrect Reasoning in GRPO
by: Chen, Peter, et al.
Published: (2025)

nvBench 2.0: Resolving Ambiguity in Text-to-Visualization through Stepwise Reasoning
by: Luo, Tianqi, et al.
Published: (2025)

Dynamic Reasoning Chains through Depth-Specialized Mixture-of-Experts in Transformer Architectures
by: Roy, Sampurna, et al.
Published: (2025)

Offline Reinforcement Learning for LLM Multi-Step Reasoning
by: Wang, Huaijie, et al.
Published: (2024)

Discovering Process-Outcome Credit in Multi-Step LLM Reasoning
by: Wang, Xiangwei, et al.
Published: (2026)

MathFimer: Enhancing Mathematical Reasoning by Expanding Reasoning Steps through Fill-in-the-Middle Task
by: Yan, Yuchen, et al.
Published: (2025)

Reinforcement Learning for Diffusion LLMs with Entropy-Guided Step Selection and Stepwise Advantages
by: Kunde, Vishnu Teja, et al.
Published: (2026)

Towards Stepwise Domain Knowledge-Driven Reasoning Optimization and Reflection Improvement
by: Liu, Chengyuan, et al.
Published: (2025)

STAR: Stepwise Task Augmentation with Relation Learning for Aspect Sentiment Quad Prediction
by: Lai, Wenna, et al.
Published: (2025)

Stepwise Penalization for Length-Efficient Chain-of-Thought Reasoning
by: Li, Xintong, et al.
Published: (2026)

ReasonBENCH: Benchmarking the (In)Stability of LLM Reasoning
by: Potamitis, Nearchos, et al.
Published: (2025)

Multilingual Performance Biases of Large Language Models in Education
by: Gupta, Vansh, et al.
Published: (2025)

IndicMMLU-Pro: Benchmarking Indic Large Language Models on Multi-Task Language Understanding
by: KJ, Sankalp, et al.
Published: (2025)

Synthetic Data Generation & Multi-Step RL for Reasoning & Tool Use
by: Goldie, Anna, et al.
Published: (2025)

LLM Reasoners: New Evaluation, Library, and Analysis of Step-by-Step Reasoning with Large Language Models
by: Hao, Shibo, et al.
Published: (2024)

MorphNLI: A Stepwise Approach to Natural Language Inference Using Text Morphing
by: Negru, Vlad Andrei, et al.
Published: (2025)

Confidence-Guided Stepwise Model Routing for Cost-Efficient Reasoning
by: Lee, Sangmook, et al.
Published: (2025)

PDTrim: Targeted Pruning for Prefill-Decode Disaggregation in Inference
by: Zhang, Hao, et al.
Published: (2025)