Theme of inquiry

How do training dynamics and architectural choices shape learned representations?

A question within its area, explored through 9 lines of inquiry below — each a family of specific questions the research asks.


What causes collapse and instability in reinforcement learning policy-critic training?

27 specific questions

See all 27 questions in this line of inquiry
Can recurrent computation achieve reasoning capabilities that fixed-depth models cannot?

34 specific questions

See all 34 questions in this line of inquiry
How does training on self-generated data affect model capabilities?

56 specific questions

See all 56 questions in this line of inquiry
Can transformers overcome architectural limits on sequential and recursive reasoning?

37 specific questions

See all 37 questions in this line of inquiry
Why do sparse parameter updates enable efficient adaptation without full retraining?

58 specific questions

See all 58 questions in this line of inquiry
How does scaling enable compositional generalization in neural networks?

11 specific questions

See all 11 questions in this line of inquiry
How does neural representation structure affect interpretability and generalization capabilities?

49 specific questions

See all 49 questions in this line of inquiry
What organizational structures emerge in learned representations?

48 specific questions

See all 48 questions in this line of inquiry
How do diffusion and autoregressive language models differ structurally and functionally?

33 specific questions

See all 33 questions in this line of inquiry