WHALE: Jointly Optimizing LLM Weights and Agent Harness

Researchers introduced WHALE, a recipe that alternately optimizes LLM weights and the agent harness, improving accuracy by 4-24 percentage points.

2026-09-03 ~ 2026-09-04 · 2 related posts

• WHALE Paper: Alternating Weight and Agent Harness Optimization Gains Up to 24 Points — Kangwook_Lee · 2026-09-03
• WHALE: a simple recipe to jointly optimize an LLM's weights and harness — lateinteraction · 2026-09-04