Rebeauty Atlas는 무엇인가요?

Rebeauty Atlas는 AI·디자인·마케팅·경제·의료 등 다양한 주제의 Medium 아티클을 수집해 태그·트랙·위키·인사이트로 탐색할 수 있게 정리한 리서치 아카이브입니다.

글은 어떻게 모이나요?

Rebeauty Atlas는 Medium에 공개된 아티클을 자동 수집해 AI·디자인·마케팅·경제·의료 등 주제별 태그·트랙·위키로 정리하는 리서치 아카이브입니다.

사이트는 어떤 섹션으로 구성되어 있나요?

Articles · Wiki · Insights · Tags 섹션으로 구성되어 있으며, PHP MVC · MySQL · shadcn-style UI · Wiki via keyword rules 로 만들어졌습니다.

Articles · Rebeauty Atlas

TCH MDA AI

SA Sanket Thakur

When RLVR Starts Cheating: Lessons from Long-Horizon GRPO Training on GSM8K

I kept training an RLVR model long enough … and discovered a new language.

Jul 06, 2026 · 8 min read · 0 claps #reward-hacking #hallucinations #llm

When RLVR Starts Cheating: Lessons from Long-Horizon GRPO Training on GSM8K

HUM MDA AI

SA Sanket Thakur

Implementing RLVR from Scratch: Training Qwen2.5 with REINFORCE on GSM8K

Building a minimal RLVR pipeline from scratch, understanding sequence log-probabilities, reward normalization, and training a reasoning…

Jun 15, 2026 · 8 min read · 0 claps #rlvr #llm #qwen

Implementing RLVR from Scratch: Training Qwen2.5 with REINFORCE on GSM8K

AI HUM MDA

EI eiko watanabe

Why Each AI Feels Rude in Its Own Way Paert2

Why does AI make excuses?

Jun 09, 2026 · 2 min read · 0 claps #reinforcement-learning #rlhf #rlaif

SPT HUM AI MDA

YO YouShin kim

Microsoft Research & KAIST — Rebellious Student: Reversed Teacher Signals를 통한 추론 탐사

1. Executive Summary & Background

May 27, 2026 · 7 min read · 0 claps #large-language-models #reinforcement-learning #rlvr

AI HUM MDA SCI

VI VinayDhoni

Cognitive QoS : A Truth-Flexibility Control Plane for Reliable and Adaptive LLM Systems

RLVR to RLCVR

Apr 23, 2026 · 11 min read · 1 claps #rlvr #ai #computer-science

Cognitive QoS : A Truth-Flexibility Control Plane for Reliable and Adaptive LLM Systems

AI HUM MDA

DA Danilo Canivel

RLVR: A Research Journey — Part 05: Better Teachers, Claude vs Gemini

Can you use AI to teach AI? Yes — but the choice of teacher matters more than you’d think.

Mar 11, 2026 · 3 min read · 3 claps #rlvr #reinforcement-learning #artificial-intelligence

RLVR: A Research Journey — Part 05: Better Teachers, Claude vs Gemini

AI HUM MDA

CI Ciprian Ghetau

Lagniappe #55: RLVR

RLVR, sau Reinforcement Learning with Verifiable Rewards (Învățare prin întărire din recompense verificabile), reprezintă următoarea…

Feb 19, 2026 · 2 min read · 0 claps #ai-agent #ai #rlvr

AI LIF MDA

TO Tobias Norlund

Reinforcement Fine-Tuning for Computer Use: Part 1 — Setting up a Scalable CUA Gym

Do you wish to build a Computer Use Agent (CUA) for certain tasks or workflows, but struggle to get it to work reliably with prompt…

Jan 18, 2026 · 8 min read · 5 claps #computer-use #rlvr #fine-tuning

Reinforcement Fine-Tuning for Computer Use: Part 1 — Setting up a Scalable CUA Gym

AI HUM MDA

GA Gaurav Garkoti

Reinforcement Learning with Verifiable Rewards (RLVR): Training LLMs for Real Reasoning

Introduction

Jan 09, 2026 · 4 min read · 1 claps paywalled #rlvr #machine-learning #reinforcement-learning

Reinforcement Learning with Verifiable Rewards (RLVR): Training LLMs for Real Reasoning

AI HUM MDA

RI Rinn

RL100-Beyond 100% Success Rate in Robotic Manipulation by Huazhe Xu-Tsinghua University & GALAXEA

In the field of robotic manipulation, existing supervised learning paradigms face inherent limitations: they often overfit to limited…

Jan 03, 2026 · 1 min read · 0 claps #embodied-ai #ai #foundation-models

RL100-Beyond 100% Success Rate in Robotic Manipulation by Huazhe Xu-Tsinghua University & GALAXEA

SOC HUM MDA AI

LJ L.J.

Can NeurIPS 25 oral RLVR really improve reasoning ability?

Sharing a NeurIPS 2025 Oral Paper. A Collaboration between Tsinghua University and Shanghai Jiao Tong University

Dec 18, 2025 · 2 min read · 5 claps #neurips #reinforcement-learning #rlvr

Can NeurIPS 25 oral RLVR really improve reasoning ability?

TCH HUM MDA AI SOC

DI Dixon

Evolving Language Models without Labels: Majority Drives Selection, Novelty Promotes Variation

Teaching LLMs to improve themselves without teachers

Sep 20, 2025 · 2 min read · 0 claps #ttrl #rlvr #evol-rl

Evolving Language Models without Labels: Majority Drives Selection, Novelty Promotes Variation

AI LIF MDA

FA Fatemeh Razgaleh

Open Training Recipes for Reasoning in Language Models

Instruction Tuning, Preference Tuning, and RVLR

May 31, 2025 · 3 min read · 0 claps #llm-agent #instruction-tuning #preference-tuning