[otabdeveloper4]

By "rigorous" I mean peeking under the curtain and actually quantifying the interactions between different system prompts and model weights.

"Chain of thought" and "reasoning" is marketing bullshit.

[int_19h]

How would you quantify it? The LM is still a black box, we don't know what most of those weights actually do.

[otabdeveloper4]

Yes, that was my question.