Author
Michael Pradel
Papers that credit Michael Pradel in their source bylines. Matching names may refer to different researchers.
2 results
BenchMeUp turns research specifications into runnable benchmarks
Across four software tasks, BenchMeUp instances more often met source, completeness, buildability, and outcome checks than coding-agent instances.
Deleting agent history works best when recent tool results stay
A study of coding agents finds delayed history deletion often preserves task solving, while estimated costs depend on cache pricing.
