Cover Image

PIKA: Expert-Level Synthetic Datasets for Post-Training Alignment from Scratch

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Yin, Shangjian, Liang, Shining, Ding, Wenbiao, Qian, Yuli, Shi, Zhouxing, Li, Hongzhi, Xie, Yutao
Format:	Preprint
Published:	2025
Subjects:	Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

GRLO: Towards Generalizable Reinforcement Learning in Open-Ended Environments from Zero
by: Yin, Shangjian, et al.
Published: (2026)

ConPress: Learning Efficient Reasoning from Multi-Question Contextual Pressure
by: Deng, Jie, et al.
Published: (2026)

Beyond Rejection Sampling: Trajectory Fusion for Scaling Mathematical Reasoning
by: Deng, Jie, et al.
Published: (2026)

Selected Languages are All You Need for Cross-lingual Truthfulness Transfer
by: Liu, Weihao, et al.
Published: (2024)

Aligning Large Language Models via Fully Self-Synthetic Data
by: Yin, Shangjian, et al.
Published: (2025)

MuDAF: Long-Context Multi-Document Attention Focusing through Contrastive Learning on Attention Heads
by: Liu, Weihao, et al.
Published: (2025)

Quantifying and Improving the Robustness of Retrieval-Augmented Language Models Against Spurious Features in Grounding Data
by: Yang, Shiping, et al.
Published: (2025)

ECLM: Entity Level Language Model for Spoken Language Understanding with Chain of Intent
by: Yin, Shangjian, et al.
Published: (2024)

EAQuant: Enhancing Post-Training Quantization for MoE Models via Expert-Aware Optimization
by: Fu, Zhongqian, et al.
Published: (2025)

Unleashing LLM Reasoning Capability via Scalable Question Synthesis from Scratch
by: Ding, Yuyang, et al.
Published: (2024)

Synthetic Data (Almost) from Scratch: Generalized Instruction Tuning for Language Models
by: Li, Haoran, et al.
Published: (2024)

TF3-RO-50M: Training Compact Romanian Language Models from Scratch on Synthetic Moral Microfiction
by: Nadas, Mihai Dan, et al.
Published: (2026)

Conan-Embedding-v2: Training an LLM from Scratch for Text Embeddings
by: Li, Shiyu, et al.
Published: (2025)

I-SHEEP: Self-Alignment of LLM from Scratch through an Iterative Self-Enhancement Paradigm
by: Liang, Yiming, et al.
Published: (2024)

Defending LLMs against Jailbreaking Attacks via Backtranslation
by: Wang, Yihan, et al.
Published: (2024)

Prompt-Level Reward Specifications for Open-Ended Post-Training
by: Weng, Zijun, et al.
Published: (2026)

Privasis: Synthesizing the Largest "Public" Private Dataset from Scratch
by: Kim, Hyunwoo, et al.
Published: (2026)

IterAlign: Iterative Constitutional Alignment of Large Language Models
by: Chen, Xiusi, et al.
Published: (2024)

TagRAG: Tag-guided Hierarchical Knowledge Graph Retrieval-Augmented Generation
by: Tao, Wenbiao, et al.
Published: (2025)

Lost in Stories: Consistency Bugs in Long Story Generation by LLMs
by: Li, Junjie, et al.
Published: (2026)

Are Expert-Level Language Models Expert-Level Annotators?
by: Tseng, Yu-Min, et al.
Published: (2024)

Measuring Diversity in Synthetic Datasets
by: Zhu, Yuchang, et al.
Published: (2025)

HART: Data-Driven Hallucination Attribution and Evidence-Based Tracing for Large Language Models
by: Liang, Shize, et al.
Published: (2026)

Training Matryoshka Mixture-of-Experts for Elastic Inference-Time Expert Utilization
by: Wang, Yaoxiang, et al.
Published: (2025)

Magpie: Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothing
by: Xu, Zhangchen, et al.
Published: (2024)

Aligning Large Language Models to a Domain-specific Graph Database for NL2GQL
by: Liang, Yuanyuan, et al.
Published: (2024)

PEO: Improving Bi-Factorial Preference Alignment with Post-Training Policy Extrapolation
by: Liu, Yuxuan
Published: (2025)

Breaking the Script Barrier in Multilingual Pre-Trained Language Models with Transliteration-Based Post-Training Alignment
by: Xhelili, Orgest, et al.
Published: (2024)

Training LayoutLM from Scratch for Efficient Named-Entity Recognition in the Insurance Domain
by: Uthayasooriyar, Benno, et al.
Published: (2024)

Post-Trained MoE Can Skip Half Experts via Self-Distillation
by: Lv, Xingtai, et al.
Published: (2026)

Constructing Multimodal Datasets from Scratch for Rapid Development of a Japanese Visual Language Model
by: Sasagawa, Keito, et al.
Published: (2024)

DiscoX: Benchmarking Discourse-Level Translation task in Expert Domains
by: Zhao, Xiying, et al.
Published: (2025)

Reasoning Pattern Alignment Merging for Adaptive Reasoning
by: Zhong, Zhaofeng, et al.
Published: (2026)

Recovering Diversity Without Losing Alignment: A DPO Recipe for Post-Trained LLMs
by: Samuel, Vinay, et al.
Published: (2026)

Why Does Reinforcement Learning Generalize? A Feature-Level Mechanistic Study of Post-Training in Large Language Models
by: Shi, Dan, et al.
Published: (2026)

Mecellem Models: Turkish Models Trained from Scratch and Continually Pre-trained for the Legal Domain
by: Uğur, Özgür, et al.
Published: (2026)

Value Drifts: Tracing Value Alignment During LLM Post-Training
by: Bhatia, Mehar, et al.
Published: (2025)

QuantMoE-Bench: Examining Post-Training Quantization for Mixture-of-Experts
by: Li, Pingzhi, et al.
Published: (2024)

XPERT: Expert Knowledge Transfer for Effective Training of Language Models
by: Liu, Chang, et al.
Published: (2026)

Generalizable LLM Learning of Graph Synthetic Data with Post-training Alignment
by: Zhang, Yizhuo, et al.
Published: (2025)