Launching today
debate.tellodb
Stop Trusting Single-LLM Bias
6 followers
Stop Trusting Single-LLM Bias
6 followers
Stop trusting single-LLM bias. debate.tellodb routes high-stakes queries through an adversarial review chain to reach consensus.
6 followers
6 followers
I have been trying this concept around and passing one LLM response to another LLM forces it to be more critical and "think out of the box". I've built a product around it, not sure about cost to token ratio for now, I'll actively listen to feedback to improve it. Lemme know what you think it has 2 free round of usage. They go round and round passing arguments until they converge or we run out of rounds to talk.