geri dön
// WRITING · 2026.03.02

Evaluating llms on quantitative tasks

benchmarks lie. building small, honest evals for math-heavy and finance-heavy workloads.

Public benchmarks are optimized against the moment they are published. By the time a model tops a leaderboard, that leaderboard has leaked into training data. For anything quantitative, you have to build your own evals — small, private, and honest.

Why generic benchmarks fail here

A model that scores well on grade-school math can still fail on:

Building an honest eval

Benchmarks tell you who is winning a game. Evals tell you whether your system works. Only one of those pays the bills.