This repository contains the data for ThrowBench. Data is given in JSONL format. Records have the following fields:
bug_id
: The RunBugRun bug idcode
: Program codeexception_type
: The name of the exception thrown, orno_exception
if none thrown. This is the target label.exception_message
: The full message of the exception thrown`language
: The langugue, eitherc_sharp
,java
,python
orruby
input
: The program input that triggers the exceptioninputs
: Other triggering inputs (input
was randomly selected from this list)locs
: Program length (in lines of code)
Model outputs can be found in the results directory. Files have the following fields:
model
: Model namepredicted_exception
: Answer given by modelactual_exception
: Ground-truth answerbug_id
: See aboveoutput
: Full model outputlanguage
: See above
The evaluation script can be found in run.py