What is Post-Training? Why many models really widen the gap is post-training
Post-training refers to the process by which a model continues to become more useful, stable, and in line with the target task through additional trai...
Admin •
31
What is Grounding? Why more and more AI searches are emphasizing "bringing source answers"
Grounding can generally be understood as "making the model's answers as based on verifiable external information as possible". The most common manifes...
Admin •
53
What is Prompt Caching? Why long prompts and multiple rounds of agents are trying to find a way to eat the cache
Prompt caching refers to caching the recurring prompt prefixes in the model request, and then trying to reuse the processing results directly when enc...
Admin •
54
What is Agentic Search? Why search products are starting to shift from "give answers" to "check for you"
Agentic Search generally refers to a new generation of search methods with the ability to "find autonomously, disassemble tasks, continuously browse a...
Admin •
59
What is a Voice Agent? Why AI voice assistants are starting to move from "talking" to "doing"
Voice Agent can be understood as "Agent with voice as the main entrance". It doesn't just turn your words into text and then read out the model reply,...
Admin •
57
What is Context Caching? Why it's becoming a cost keyword for long-context products
Context caching refers to caching a context that will be sent to the model repeatedly and reusing subsequent requests as much as possible, rather than...
Admin •
54