A new GEPA visualizer shows an LLM optimizing its own prompts in real time.

Most prompt optimizers treat the LLM like a black box.

They reduce every run to a single score and call it learning.

GEPA takes a different path.

It reads the full execution trace, error logs, and… pic.twitter.com/ixJEsIMutd

— AlphaSignal AI (@AlphaSignalAI) 31 mai 2026

A new GEPA visualizer shows an LLM optimizing its own prompts in real time. Most prompt optimizers treat the LLM like a black box. They reduce every run to a single score and call it learning. GEPA takes a different path. It reads the full execution trace, error logs, and

→ View original post on X — @alphasignalai