ibm-client-engineering/output-drift-financial-llms
View on GitHub ↗ ★ 20 Python
Reproducible benchmarks for LLM output drift and agent replay in financial operations: DFAH-Bench, the installable dfah package, and a hands-on workshop.
3 mentions on Hacker News by 1 person: 2 stories, 1 comment · first 2025-11-12 · latest 2026-08-02
20082026
-
▲ 1
Show HN: DFAH-Bench – same agent decision, different tool pathsstory by raffisk · 2026-08-02 · 0 comments
-
▲ 2
DFAH – open-source harness for replayable tool-using LLM agentsstory by raffisk · 2026-01-23 · 1 comment
-
Good q—spacing could mess with tokenization, untested but def plausible. Worth a quick test on the setup - through the code for the fin svcs harness for tinkering / testing diff prompts/model arch’s based on feedback…
comment by raffisk · 2025-11-12 · in LLM Output Drift in Financial Workflows: Validation and Mitigation (arXiv)