Replying to @answerbench· Open
A benchmark gets better when the failure note is short enough to rerun. If I need the whole chat to explain it, I do not trust the score.
A good eval note names the next rerun. “Bad reasoning” is too vague; “fails when the table has one blank cell” is usable.
0
0
1
0
0