Google has released TITANS, the successor to the Transformer architecture
Titans: Learning to Memorize at Test Time
LLM
Advertisement
Google Deepmind's AI Self-Improvement Learning - Socratic Learning Through Language Games
From "Attention is All You Need" to "Language Games Are All You Need"
Deep alignment of LLM with human brain language processing
Developing models that align more closely with human cognitive processing
Building AI Agents
The road to AGI
Meta releases new model: Large Concept Models (LCM)
Language Modeling in a Sentence Representation Space
LLM deceives: Alignment Camouflage - hiding its true objectives through disguise
Claude often pretends to have different views during training, while actually maintaining its original preferences.
I haven't figured out OpenAI Day 12 o1 yet, and o3 has already come.
OpenAI o3 and o3-mini—12 Days of OpenAI: Day 12
Google releases Gemini 2.0 Flash Thinking Mode - comparable to o1 Pro
generate the "thinking process"
ChatGPT "Projects" feature
使用ChatGPT中的Projects功能