Skip to content

benchmark

Context Rot report

Chroma research report measuring how language model accuracy varies with input length across 18 models, including tasks simple enough that length alone should not matter.

Known aliases

  • Context Rot

Relationships

No evidence-backed relationships are recorded.

Current clusters