[P] asmsense: 228k stripped x86-64 functions with source-aware labels, and a map of the data we still lack. [P]

Explain Like I'm Normal

A new dataset and model called asmsense provides 228k labeled x86-64 functions to help AI identify and describe code in stripped binaries. By compiling open-source C projects with original metadata, it allows LLMs like Qwen to map assembly back to human-readable intent. This bridges a major gap in automated reverse engineering and security analysis.