Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Matlin, Glenn, Siddharth, JM, Anirudh, Shukla, Aditya, Hassan, Yahya, Chava, Sudheer
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2512.08965
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917135763111936
author Matlin, Glenn
Siddharth
JM, Anirudh
Shukla, Aditya
Hassan, Yahya
Chava, Sudheer
author_facet Matlin, Glenn
Siddharth
JM, Anirudh
Shukla, Aditya
Hassan, Yahya
Chava, Sudheer
contents Language Models (LMs) struggle with complex, interdependent instructions, particularly in high-stakes domains like finance where precision is critical. We introduce FIFE, a novel, high-difficulty benchmark designed to assess LM instruction-following capabilities for financial analysis tasks. FIFE comprises 88 human-authored prompts and employs a verification system with chainable, verifiable constraints for fine-grained reward signals. We evaluate 53 models (proprietary, open-weight, open-source) in a zero-shot setting. Our key findings reveal a clear performance hierarchy: the top open-weight model (76.1 strict / 79.5 loose) surpasses the leading proprietary system (65.9 strict / 70.5 loose), while the best open-source models lag significantly (45.5 strict / 48.9 loose). However, even top-performing models struggle with FIFE's complex requirements, failing to achieve perfect compliance. We release our dataset and code as an open-source resource to promote research in Reinforcement Learning for the financial domain.
format Preprint
id arxiv_https___arxiv_org_abs_2512_08965
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Financial Instruction Following Evaluation (FIFE)
Matlin, Glenn
Siddharth
JM, Anirudh
Shukla, Aditya
Hassan, Yahya
Chava, Sudheer
Machine Learning
Artificial Intelligence
Computation and Language
Language Models (LMs) struggle with complex, interdependent instructions, particularly in high-stakes domains like finance where precision is critical. We introduce FIFE, a novel, high-difficulty benchmark designed to assess LM instruction-following capabilities for financial analysis tasks. FIFE comprises 88 human-authored prompts and employs a verification system with chainable, verifiable constraints for fine-grained reward signals. We evaluate 53 models (proprietary, open-weight, open-source) in a zero-shot setting. Our key findings reveal a clear performance hierarchy: the top open-weight model (76.1 strict / 79.5 loose) surpasses the leading proprietary system (65.9 strict / 70.5 loose), while the best open-source models lag significantly (45.5 strict / 48.9 loose). However, even top-performing models struggle with FIFE's complex requirements, failing to achieve perfect compliance. We release our dataset and code as an open-source resource to promote research in Reinforcement Learning for the financial domain.
title Financial Instruction Following Evaluation (FIFE)
topic Machine Learning
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2512.08965