·
DataBubble
  • Home
  • Models
  • News
  • Compare
  • Boards
  • Pricing
  • About
  • Newsletter
  • Methodology
  • Contact
Latest
NewsBench: Expert-Grounded Evaluation of Epistemic Quality in AI News Reporting6h◆Foundations of the Scaling Law6h◆AdaPRA: Process Reward Agents with Adaptive Rubrics for Data Science6h◆Scoped Semantic Invariance as a Foundation for LLM Reliability6h◆An actively trained neural operator for near-explicit atmospheric chemistry6h◆Momentum in large-batch training: Polyak enlarges the critical batch size, Nesterov improves data efficiency6h◆InferPriv: Evaluating Inferential Privacy Leakage in LLM Assistants6h◆ThinkBit: Search Codewords, Recalibrate Scales, Select by Reasoning6h◆Multimodal Requirements Traceability Link Recovery in Industrial Settings6h◆The Price of Scale in Proactive Memory Recall6h◆The Watchtower Imperative: Lifespan-Calibrated Evaluation for Frontier AI6h◆The Value of Memory in Language Agents6h◆HyScale: Scaling Long-Context Training for Learned Sparse Attention6h◆Training an MoE LLM in MXFP4: How We Closed the Gap to BF166h◆CodeJeNN: A simple C++ neural network generator for physics applications6h◆WiMoT: Replay-Free Continual Wi-Fi Sensing via Moment Transport6h◆I-HITL: Inalienable Human-in-the-Loop6h◆Advancing SVD-based LLM Compression via Layer-Wise Error Model Search6h◆Discovering and Reasoning of Causality in the Hidden World with Large Language Models6h◆DeALOG: Decentralized Multi-Agents Log-Mediated Reasoning Framework6h◆NewsBench: Expert-Grounded Evaluation of Epistemic Quality in AI News Reporting6h◆Foundations of the Scaling Law6h◆AdaPRA: Process Reward Agents with Adaptive Rubrics for Data Science6h◆Scoped Semantic Invariance as a Foundation for LLM Reliability6h◆An actively trained neural operator for near-explicit atmospheric chemistry6h◆Momentum in large-batch training: Polyak enlarges the critical batch size, Nesterov improves data efficiency6h◆InferPriv: Evaluating Inferential Privacy Leakage in LLM Assistants6h◆ThinkBit: Search Codewords, Recalibrate Scales, Select by Reasoning6h◆Multimodal Requirements Traceability Link Recovery in Industrial Settings6h◆The Price of Scale in Proactive Memory Recall6h◆The Watchtower Imperative: Lifespan-Calibrated Evaluation for Frontier AI6h◆The Value of Memory in Language Agents6h◆HyScale: Scaling Long-Context Training for Learned Sparse Attention6h◆Training an MoE LLM in MXFP4: How We Closed the Gap to BF166h◆CodeJeNN: A simple C++ neural network generator for physics applications6h◆WiMoT: Replay-Free Continual Wi-Fi Sensing via Moment Transport6h◆I-HITL: Inalienable Human-in-the-Loop6h◆Advancing SVD-based LLM Compression via Layer-Wise Error Model Search6h◆Discovering and Reasoning of Causality in the Hidden World with Large Language Models6h◆DeALOG: Decentralized Multi-Agents Log-Mediated Reasoning Framework6h◆
News/GFT: From Imitation to Reward Fine-Tuning with Unbiased Group Advantages and Dynamic Coefficient Rectification
arxiv
PublishedApril 18, 2026 at 4:00 AM
—neutral

GFT: From Imitation to Reward Fine-Tuning with Unbiased Group Advantages and Dynamic Coefficient Rectification

Source
arxiv.orgfull article ↗
Read on arxiv→
Publisher summary· verbatim

arXiv:2604.14258v1 Announce Type: cross Abstract: Large language models are typically post-trained using supervised fine-tuning (SFT) and reinforcement learning (RL), yet effectively unifying efficient knowledge injection with robust generalization remains challenging. In this work, we provide a tra

Stay posted· Newsletter

Three short emails a week: Monday, Wednesday and Friday. Top movers, new models and the most-referenced story.

// no spam · unsubscribe one-click · free forever

Your email is used only to send The Bubble Brief. Privacy policy

Discussion
Source
↗
arxiv
Read original ↗All from arxiv →
Tags
03
#language-models#fine-tuning#reinforcement-learning

No replies yet. Be first.

Source
↗
arxiv
Read original ↗All from arxiv →
Tags
03
#language-models#fine-tuning#reinforcement-learning

Related coverage

More from ARXIV
arxivNewsBench: Expert-Grounded Evaluation of Epistemic Quality in AI News Reporting6harxivFoundations of the Scaling Law6harxivAdaPRA: Process Reward Agents with Adaptive Rubrics for Data Science6harxivScoped Semantic Invariance as a Foundation for LLM Reliability6h
The Bubble Brief
MON · WED · FRI

Read language-models insights every Monday, Wednesday and Friday — top movers, new models, story of the week.

// no spam · unsubscribe one-click · free forever

Your email is used only to send The Bubble Brief. Privacy policy

Originally published on arxiv ↗
PrivacyTermsRefundsLegal notice|AboutMethodologyContact

© 2026 Databubble

Built by Marouane Gazouzi · Readyo, AI interview coach

HomeModelsNews