Model evaluation on minif2f fails?

When I evaluate InternLM2-Math-Plus-7b in minif2f through this code, it fails. The model only generates one line "Here is the predicted next tactic:" without any tactics. If I let the model continue generating until I get a tactic each time, I only get a pass rate 38.1% instead of 43.4%.

Start by reproducing the InternLM2-Math-Plus-7b evaluation through the repository's minif2f evaluation code and inspect why generation stops after "Here is the predicted next tactic:". Compare the behavior with the reported 43.4% and 38.1% pass rates; done means the failure is explained and the evaluation produces the intended tactic output or a documented limitation.

Escrito por el modelo de indexación a partir del texto del issue.