{"id":50670358,"url":"https://github.com/9bow/learn-rl-for-llms","last_synced_at":"2026-06-08T10:33:43.906Z","repository":{"id":351230008,"uuid":"1209234584","full_name":"9bow/learn-rl-for-llms","owner":"9bow","description":"LLM을 위한 강화학습 종합 학습 가이드 - RL for LLMs Learning Site","archived":false,"fork":false,"pushed_at":"2026-04-14T05:14:51.000Z","size":476,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-04-14T07:25:30.532Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"MDX","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/9bow.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-04-13T08:16:45.000Z","updated_at":"2026-04-14T05:14:55.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/9bow/learn-rl-for-llms","commit_stats":null,"previous_names":["9bow/learn-rl-for-llms"],"tags_count":null,"template":false,"template_full_name":null,"purl":"pkg:github/9bow/learn-rl-for-llms","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/9bow%2Flearn-rl-for-llms","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/9bow%2Flearn-rl-for-llms/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/9bow%2Flearn-rl-for-llms/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/9bow%2Flearn-rl-for-llms/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/9bow","download_url":"https://codeload.github.com/9bow/learn-rl-for-llms/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/9bow%2Flearn-rl-for-llms/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":34059156,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-06-08T02:00:07.615Z","response_time":111,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2026-06-08T10:33:42.750Z","updated_at":"2026-06-08T10:33:43.901Z","avatar_url":"https://github.com/9bow.png","language":"MDX","funding_links":[],"categories":[],"sub_categories":[],"readme":"# LLM을 위한 강화학습\n\nLLM 분야에서 쓰이는 강화학습 기법을 기초부터 RLHF, DPO, GRPO, RLVR, 코드 RL, Vision RL, 에이전트 RL, 인프라와 안전성까지 연결해 설명하는 한국어 학습 사이트입니다.\n\n## 대상\n\n- 강화학습 기본 개념을 LLM 맥락에서 다시 정리하려는 연구자와 엔지니어\n- RLHF, DPO, GRPO, RLVR, PRM 같은 정렬·추론 학습 기법의 차이를 알고 싶은 학습자\n- LLM 학습 파이프라인, 평가, 분산 RL 인프라를 함께 이해하려는 팀\n\n## 구성\n\n- 11개 섹션\n- 61개 MDX 챕터\n- 11개 섹션별 퀴즈\n- 문서 품질 점검 스크립트 포함\n- Astro 6 + Starlight + React 기반\n- 수식 렌더링을 위한 KaTeX 설정 포함\n\n## 커리큘럼\n\n1. 강화학습 기초: MDP, 정책, 가치 함수, policy gradient, PPO, 보상 설계\n2. LLM과 RL의 만남: LLM을 RL 에이전트로 보는 관점, KL penalty, 온라인/오프라인 RL\n3. RLHF: 보상 모델, PPO fine-tuning, 한계, Constitutional AI, InstructGPT 사례\n4. 직접 선호 최적화: DPO, IPO, KTO, ORPO, SimPO, RainbowPO\n5. Reasoning과 RL: GRPO, DAPO, PRM, STaR, Quiet-STaR, 탐색과 test-time scaling\n6. RLVR: 검증 가능한 보상, 수학/코드 RLVR, DeepSeek-R1, rule-based reward, scaling\n7. 코드 생성과 RL: 코드 RL 기본, StepCoder, process reward, AceMath, 오픈소스 추론 모델\n8. Vision 모델과 RL: 멀티모달 RLHF, RLSD, diffusion RL, reward model 설계\n9. 에이전트 RL: 웹 에이전트, tool calling, multi-turn, multi-agent RL\n10. RL 인프라: TRL, OpenRLHF, veRL, 분산 학습, 실험 관리\n11. 안전성과 미래: 안전 정렬, scalable oversight, 보상 해킹, 열린 문제\n\n## 로컬 개발\n\n```bash\npnpm install\npnpm dev\npnpm build\npnpm preview\n```\n\n## 문서 검증\n\n```bash\npnpm check:docs\npnpm verify:scripts\n```\n\n## 주요 경로\n\n- 문서 본문: `src/content/docs/`\n- 퀴즈 데이터: `public/data/quiz/`\n- 문서 검증 스크립트: `scripts/`\n- 공통 학습 컴포넌트: `src/components/learning/`\n- 사이트 설정: `astro.config.mjs`\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2F9bow%2Flearn-rl-for-llms","html_url":"https://awesome.ecosyste.ms/projects/github.com%2F9bow%2Flearn-rl-for-llms","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2F9bow%2Flearn-rl-for-llms/lists"}