Skip to content

Latest commit

 

History

History
13 lines (9 loc) · 590 Bytes

README.md

File metadata and controls

13 lines (9 loc) · 590 Bytes

r2ai-model

Collection of data sources to generate a dataset for training and finetuning LLM models to use radare2.

Organization

Dataset is stored in Q/A form (Question/Answer) separating them by tabs (TSV) where the question is phrased in English and the answer is an r2 oneliner to be executed by r2ai in auto mode.

  • / -> root directory, scripts to generate raw QA
  • data/radare2_ok.tsv -> validated statements
  • data/radare2_todo.tsv -> unanswered questions
  • data/Attic/ -> already processed files
  • data/sources -> unfiltered data sources to be used to generate questions