Cover Image

The Signal is in the Steps: Local Scoring for Reasoning Data Selection

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Just, Hoang Anh, Ko, Myeongseob, Jia, Ruoxi
Format:	Preprint
Published:	2025
Subjects:	Machine Learning Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Probing Knowledge Holes in Unlearned LLMs
by: Ko, Myeongseob, et al.
Published: (2025)

Characterizing Model-Native Skills
by: Kang, Feiyang, et al.
Published: (2026)

DiPT: Enhancing LLM reasoning through diversified perspective-taking
by: Just, Hoang Anh, et al.
Published: (2024)

Get more for less: Principled Data Selection for Warming Up Fine-Tuning in LLMs
by: Kang, Feiyang, et al.
Published: (2024)

Boosting Alignment for Post-Unlearning Text-to-Image Generative Models
by: Ko, Myeongseob, et al.
Published: (2024)

Data Valuation and Selection in a Federated Model Marketplace
by: Li, Wenqian, et al.
Published: (2025)

Boosting Deductive Reasoning with Step Signals In RLHF
by: Li, Jialian, et al.
Published: (2024)

The Mirrored Influence Hypothesis: Efficient Data Influence Estimation by Harnessing Forward Passes
by: Ko, Myeongseob, et al.
Published: (2024)

Data-Centric Human Preference with Rationales for Direct Preference Alignment
by: Just, Hoang Anh, et al.
Published: (2024)

From Weak Cues to Real Identities: Evaluating Inference-Driven De-Anonymization in LLM Agents
by: Ko, Myeongseob, et al.
Published: (2026)

Injecting Measurement Information Yields a Fast and Noise-Robust Diffusion-Based Inverse Problem Solver
by: Patsenker, Jonathan, et al.
Published: (2025)

Do Latent-CoT Models Think Step-by-Step? A Mechanistic Study on Sequential Reasoning Tasks
by: Liang, Jia, et al.
Published: (2026)

Retracing the Past: LLMs Emit Training Data When They Get Lost
by: Ko, Myeongseob, et al.
Published: (2025)

LLM Reasoning as Trajectories: Step-Specific Representation Geometry and Correctness Signals
by: Sun, Lihao, et al.
Published: (2026)

Step-DPO: Step-wise Preference Optimization for Long-chain Reasoning of LLMs
by: Lai, Xin, et al.
Published: (2024)

Selective Steering: Norm-Preserving Control Through Discriminative Layer Selection
by: Dang, Quy-Anh, et al.
Published: (2026)

Efficient Bilevel Optimization for Meta Label Correction in Noisy Label Learning
by: Nguyen, Ba Hoang Anh, et al.
Published: (2026)

Deciphering Scientific Reasoning Steps from Outcome Data for Molecule Optimization
by: Liu, Zequn, et al.
Published: (2026)

Score-of-Mixture Training: Training One-Step Generative Models Made Simple via Score Estimation of Mixture Distributions
by: Jayashankar, Tejas, et al.
Published: (2025)

Quagmires in SFT-RL Post-Training: When High SFT Scores Mislead and What to Use Instead
by: Kang, Feiyang, et al.
Published: (2025)

Can Small Training Runs Reliably Guide Data Curation? Rethinking Proxy-Model Practice
by: Wang, Jiachen T., et al.
Published: (2025)

One Step Forward and K Steps Back: Better Reasoning with Denoising Recursion Models
by: Cameron, Chris, et al.
Published: (2026)

Learning to Generate Formally Verifiable Step-by-Step Logic Reasoning via Structured Formal Intermediaries
by: Chen, Luoxin, et al.
Published: (2026)

PHEATPRUNER: Interpretable Data-centric Feature Selection for Multivariate Time Series Classification through Persistent Homology
by: Pham, Anh-Duy, et al.
Published: (2025)

Variance Alignment Score: A Simple But Tough-to-Beat Data Selection Method for Multimodal Contrastive Learning
by: Wang, Yiping, et al.
Published: (2024)

LLM Reasoning with Process Rewards for Outcome-Guided Steps
by: Rezaei, Mohammad, et al.
Published: (2026)

Just Enough Shifts: Mitigating Over-Refusal in Aligned Language Models with Targeted Representation Fine-Tuning
by: Dabas, Mahavir, et al.
Published: (2025)

Do We Need to Verify Step by Step? Rethinking Process Supervision from a Theoretical Perspective
by: Jia, Zeyu, et al.
Published: (2025)

Flatness-aware Sequential Learning Generates Resilient Backdoors
by: Pham, Hoang, et al.
Published: (2024)

The Hidden Signal of Verifier Strictness: Controlling and Improving Step-Wise Verification via Selective Latent Steering
by: Zhou, Yefan, et al.
Published: (2026)

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

From Few to Many: Self-Improving Many-Shot Reasoners Through Iterative Optimization and Generation
by: Wan, Xingchen, et al.
Published: (2025)

Generative Score Inference for Multimodal Data
by: Tian, Xinyu, et al.
Published: (2026)

Integrating Score-Based Diffusion Models with Machine Learning-Enhanced Localization for Advanced Data Assimilation in Geological Carbon Storage
by: Seabra, Gabriel Serrão, et al.
Published: (2025)

Step-KTO: Optimizing Mathematical Reasoning through Stepwise Binary Feedback
by: Lin, Yen-Ting, et al.
Published: (2025)

AdaptiveStep: Automatically Dividing Reasoning Step through Model Confidence
by: Liu, Yuliang, et al.
Published: (2025)

RetroReasoner: A Reasoning LLM for Strategic Retrosynthesis Prediction
by: Ko, Hanbum, et al.
Published: (2026)

RePCS: Diagnosing Data Memorization in LLM-Powered Retrieval-Augmented Generation
by: Anh, Le Vu, et al.
Published: (2025)

Optimized Conformal Selection: Powerful Selective Inference After Conformity Score Optimization
by: Bai, Tian, et al.
Published: (2024)

AutoScale: Scale-Aware Data Mixing for Pre-Training LLMs
by: Kang, Feiyang, et al.
Published: (2024)