TY  - RPRT
TI  - From Output to Evaluation: Does Raw Instruction-Tuned Code LLMs Output Suffice for Fill-in-the-Middle Code Generation?
AU  - Wasi Uddin Ahmad
AU  - Somshubra Majumdar
AU  - Boris Ginsburg
PY  - 2025
UR  - https://arxiv.org/abs/2505.18789
ID  - 2505.18789
ER  -