Georgia Tech's 2026 LLM Course: From MoE and Self-Play RL to Diffusion LMs
cocoweixu · x · 2026-08-18
Wei Xu has updated Georgia Tech's CS 8803 Large Language Models course with a 2026 reading list organized by topic: pretraining, embeddings, MoE, reasoning and alignment, agents, long-context, attention, RL (GRPO/DAPO/R1-Zero-style training), test-time scaling, diffusion LMs, safety, and interpretability.
- Notable papers include the Chameleon data-mixing framework, DeepSeek-V2/V3.2, DPO, Attention Sink, and Absolute Zero reinforced self-play reasoning.
- Running from January 2026, with a full schedule and paper links — a solid systematic learning roadmap for LLMs.
Related event: Georgia Tech Updates LLM Course for 2026(2 posts)→
More from Research
- RareBench results: Gemini 3.7 Flash leaps ahead, DeepSeek Pro shows no gain — danielmckinn0n · 2026-08-18
- Claude AI raises proven Riemann-zeta lower bound to 67.2% — thione · 2026-08-18
- Anthropic research documents failure patterns in frontier multiagent systems — thione · 2026-08-18
- Technical discussion on LLM vocab bottlenecks: smaller vocab, longer sequences? — LChoshen · 2026-08-18
- New research: are LLM agents time-aware? Can they estimate task duration? — maksym_andr · 2026-08-18
- Paper examines moderation in AI-generated sexual content communities — chaumian · 2026-08-18