Driver metadata
- Slug
- codebert
- Status
- retractedOriginal CodeBERT WRONG verdict retracted 2026-05-13 (1000-way vs 200-way pool methodology mismatch). Driver disabled.
- Protocol match
- —Protocol-match tier not declared on the driver.
- Agent version
- v0.1.0-codebert-csn-python-mrr-3slice
- Paper
- CodeBERT: A Pre-Trained Model for Programming and Natural Languages
- Model
- microsoft/codebert-base
Source
Every driver is a TypeScript orchestration script that routes the result through the Verdict Validator. Drivers with runnable public checkpoints also link the hermetic Modal job that loads the model and measures the paper claim.
- scripts/run-reproduction-codebert.ts ↗— TS orchestrator (claim citation, DB write, validator wiring)
- modal/reproduce_codebert.py ↗— Python Modal job (HF model load, evaluation harness, measurement)