On: An empirical study of harness design for coding agents

[alansaber]

I sympathise completely but a ~3B parameter model and ~3T parameter model are going to exhibit very different behaviour, one can only infer so much large model behaviour from the former.