Buchumschlag

Can an Easy-to-Hard Curriculum Make Reasoning Emerge in Small Language Models? Evidence from a Four-Stage Curriculum on GPT-2

Fuente: arXiv

Gespeichert in:

Bibliographische Detailangaben
1. Verfasser:	Fu, Xiang
Format:	Preprint
Veröffentlicht:	2025
Schlagworte:	Computation and Language
Online-Zugang:	Acceder al recurso
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

Ähnliche Einträge

Rethinking Easy-to-Hard: Limits of Curriculum Learning in Post-Training for Deductive Reasoning
von: Mordig, Maximilian, et al.
Veröffentlicht: (2026)

Curriculum Reinforcement Learning from Easy to Hard Tasks Improves LLM Reasoning
von: Parashar, Shubham, et al.
Veröffentlicht: (2025)

What Kind of Language is Easy to Language-Model Under Curriculum Learning?
von: El-Naggar, Nadine, et al.
Veröffentlicht: (2026)

Small Language Models for Curriculum-based Guidance
von: Katharakis, Konstantinos, et al.
Veröffentlicht: (2025)

IRIS: Interleaved Reinforcement with Incremental Staged Curriculum for Cross-Lingual Mathematical Reasoning
von: Gupta, Navya, et al.
Veröffentlicht: (2026)

MERLIN: Multi-Stage Curriculum Alignment for Multilingual Encoder-LLM Integration in Cross-Lingual Reasoning
von: Uemura, Kosei, et al.
Veröffentlicht: (2025)

Infi-MMR: Curriculum-based Unlocking Multimodal Reasoning via Phased Reinforcement Learning in Multimodal Small Language Models
von: Liu, Zeyu, et al.
Veröffentlicht: (2025)

Let's Be Self-generated via Step by Step: A Curriculum Learning Approach to Automated Reasoning with Large Language Models
von: Luo, Kangyang, et al.
Veröffentlicht: (2024)

Making LLMs Better Many-to-Many Speech-to-Text Translators with Curriculum Learning
von: Du, Yexing, et al.
Veröffentlicht: (2024)

RuCL: Stratified Rubric-Based Curriculum Learning for Multimodal Large Language Model Reasoning
von: Chen, Yukun, et al.
Veröffentlicht: (2026)

TinyHelen's First Curriculum: Training and Evaluating Tiny Language Models in a Simpler Language Environment
von: Yang, Ke, et al.
Veröffentlicht: (2024)

Fast on the Easy, Deep on the Hard: Efficient Reasoning via Powered Length Penalty
von: Ling, Zehui, et al.
Veröffentlicht: (2025)

Curriculum-Guided Layer Scaling for Language Model Pretraining
von: Singh, Karanpartap, et al.
Veröffentlicht: (2025)

Reasoning Curriculum: Bootstrapping Broad LLM Reasoning from Math
von: Pang, Bo, et al.
Veröffentlicht: (2025)

FastCuRL: Curriculum Reinforcement Learning with Stage-wise Context Scaling for Efficient Training R1-like Reasoning Models
von: Song, Mingyang, et al.
Veröffentlicht: (2025)

AI "News" Content Farms Are Easy to Make and Hard to Detect: A Case Study in Italian
von: Puccetti, Giovanni, et al.
Veröffentlicht: (2024)

Training Large Language Models for Reasoning through Reverse Curriculum Reinforcement Learning
von: Xi, Zhiheng, et al.
Veröffentlicht: (2024)

LightReasoner: Can Small Language Models Teach Large Language Models Reasoning?
von: Wang, Jingyuan, et al.
Veröffentlicht: (2025)

Multi-Stage Evolutionary Model Merging with Meta Data Driven Curriculum Learning for Sentiment-Specialized Large Language Modeling
von: Inoshita, Keito, et al.
Veröffentlicht: (2026)

Can Large Language Models Automatically Jailbreak GPT-4V?
von: Wu, Yuanwei, et al.
Veröffentlicht: (2024)

Language Models Do Hard Arithmetic Tasks Easily and Hardly Do Easy Arithmetic Tasks
von: Gambardella, Andrew, et al.
Veröffentlicht: (2024)

Can Large Language Models Always Solve Easy Problems if They Can Solve Harder Ones?
von: Yang, Zhe, et al.
Veröffentlicht: (2024)

Curriculum Learning for Small Code Language Models
von: Naïr, Marwa, et al.
Veröffentlicht: (2024)

What do Language Models Learn and When? The Implicit Curriculum Hypothesis
von: Liu, Emmy, et al.
Veröffentlicht: (2026)

TACLer: Tailored Curriculum Reinforcement Learning for Efficient Reasoning
von: Lai, Huiyuan, et al.
Veröffentlicht: (2026)

DARC: Decoupled Asymmetric Reasoning Curriculum for LLM Evolution
von: Fan, Shengda, et al.
Veröffentlicht: (2026)

Beyond Scale: Small Language Models are Comparable to GPT-4 in Mental Health Understanding
von: Jia, Hong, et al.
Veröffentlicht: (2025)

SARI: Structured Audio Reasoning via Curriculum-Guided Reinforcement Learning
von: Wen, Cheng, et al.
Veröffentlicht: (2025)

Progressive Mastery: Customized Curriculum Learning with Guided Prompting for Mathematical Reasoning
von: Wu, Muling, et al.
Veröffentlicht: (2025)

Pre-Training Curriculum for Multi-Token Prediction in Language Models
von: Aynetdinov, Ansar, et al.
Veröffentlicht: (2025)

Do Syntactic Categories Help in Developmentally Motivated Curriculum Learning for Language Models?
von: Güven, Arzu Burcu, et al.
Veröffentlicht: (2025)

Less is More: Pre-Training Cross-Lingual Small-Scale Language Models with Cognitively-Plausible Curriculum Learning Strategies
von: Salhan, Suchir, et al.
Veröffentlicht: (2024)

GraphDancer: Training LLMs to Explore and Reason over Graphs via Two-Stage Curriculum Post-Training
von: Bai, Yuyang, et al.
Veröffentlicht: (2026)

It's Not Easy Being Wrong: Large Language Models Struggle with Process of Elimination Reasoning
von: Balepur, Nishant, et al.
Veröffentlicht: (2023)

Emergent Misalignment is Easy, Narrow Misalignment is Hard
von: Soligo, Anna, et al.
Veröffentlicht: (2026)

MedCLM: Learning to Localize and Reason via a CoT-Curriculum in Medical Vision-Language Models
von: Kim, Soo Yong, et al.
Veröffentlicht: (2025)

Improving Complex Reasoning over Knowledge Graph with Logic-Aware Curriculum Tuning
von: Xia, Tianle, et al.
Veröffentlicht: (2024)

Can Safety Emerge from Weak Supervision? A Systematic Analysis of Small Language Models
von: Saha, Punyajoy, et al.
Veröffentlicht: (2026)

Can Small Language Models be Good Reasoners for Sequential Recommendation?
von: Wang, Yuling, et al.
Veröffentlicht: (2024)

Mentor-KD: Making Small Language Models Better Multi-step Reasoners
von: Lee, Hojae, et al.
Veröffentlicht: (2024)