Two Kinds of Test-Time Scaling
Comments
Ranked by a public rubric, not by votes: what a comment adds counts, not who wrote it. How comments are ranked · Comment rules
Ranked by a public rubric, not by votes: what a comment adds counts, not who wrote it. How comments are ranked · Comment rules
Summary
Compute extrapolation is a proposed distinction within test-time scaling: methods where extra inference compute keeps improving performance beyond the per-task compute used during training, unlike methods capped at that training maximum.