BestHub
Sign in
Bighead's Algorithm Notes
Bighead's Algorithm Notes
Mar 19, 2026 · Artificial Intelligence

How ASDA Generates Structured Financial Reasoning Skills for LLMs Without Fine‑Tuning

The ASDA framework automatically creates modular, version‑controlled financial‑reasoning skill files by iteratively analyzing student model failures, clustering errors, and injecting structured guidance, achieving up to a 17.33‑point boost on arithmetic tasks and a 5.95‑point boost on non‑arithmetic tasks in the FAMMA benchmark, far surpassing prior zero‑training methods such as GEPA and ACE.

ASDALLM adaptationbenchmark FAMMA
0 likes · 22 min read
How ASDA Generates Structured Financial Reasoning Skills for LLMs Without Fine‑Tuning
Bighead's Algorithm Notes
Bighead's Algorithm Notes
Sep 11, 2025 · Artificial Intelligence

Fin-PRM: Alibaba’s Dianjin Team Introduces a Domain-Specific Process Reward Model for Financial Reasoning

Fin‑PRM, a domain‑specific process reward model for financial reasoning introduced by Alibaba’s Dianjin team, employs dual‑level step and trajectory rewards to provide fine‑grained supervision, achieving up to 12.9% accuracy gains in supervised fine‑tuning and 5.1% improvements in Best‑of‑N inference on benchmarks such as CFLUE and FinQA.

CFLUEFin-PRMFinQA
0 likes · 11 min read
Fin-PRM: Alibaba’s Dianjin Team Introduces a Domain-Specific Process Reward Model for Financial Reasoning