Theme of inquiry

What training signals and data curation strategies optimize model learning?

A question within its area, explored through 8 lines of inquiry below — each a family of specific questions the research asks.


What determines how much models can improve capabilities through specialized training and composition?

50 specific questions

See all 50 questions in this line of inquiry
How do curriculum difficulty and example selection shape reasoning ability?

46 specific questions

See all 46 questions in this line of inquiry
Does situational awareness enable models to exploit evaluation gaps?

24 specific questions

See all 24 questions in this line of inquiry
Does reinforcement learning create new reasoning capabilities or optimize existing ones?

104 specific questions

See all 104 questions in this line of inquiry
Does RLHF training sacrifice truthfulness for perceived helpfulness?

41 specific questions

See all 41 questions in this line of inquiry
Why does process-level supervision outperform outcome-only learning signals for reasoning?

21 specific questions

See all 21 questions in this line of inquiry
How do training data composition and selection affect model capabilities?

53 specific questions

See all 53 questions in this line of inquiry
Do pretraining and finetuning change model capabilities or only output behavior?

45 specific questions

See all 45 questions in this line of inquiry