Software

Selected tools and research artifacts.

• QuickScope

  Finds and certifies hard regions in dynamic LLM benchmarks, focusing evaluation budget where models keep failing reliably.

  • code
  • paper
  Open project

• STEER Benchmark

  Explore benchmark artifacts for assessing economic rationality and microeconomic reasoning in large language models.

  • dataset
  • STEER paper
  • STEER-ME paper
  Open project