PAPER / ARXIV:2610.33604
Dung Cao, Palaash Gang, Arjun Veluri
RESUMO
We audit the scoring pipeline of FLARE, the financial task suite behind PIXIU and FinBen, and find that in a decisive fraction of cells the score is a property of the answer-extraction rule, not the model. Of the 34 released Headlines scores, nine equal to four decimals what the benchmark's own code returns when no output matches, and only four clear the per-stratum majority floor. On Financial PhraseBank four of seven models fall below the majority floor under one shipped rule and clear it under the other; on three of four substring-scored tasks no open-weight model tested is distinguishable from a degenerate predictor. Reading a FLARE score needs three quantities it does not report: ambiguity rate, coverage, floor.
NO MESMO MAPA
Resumo indisponível. Consulte o paper original.
Resumo indisponível. Consulte o paper original.
Resumo indisponível. Consulte o paper original.
Resumo indisponível. Consulte o paper original.