Martin Görner
@martin-gorner.bsky.social
AI/ML engineer. Previously at Google: Product Manager for Keras and TensorFlow and developer advocate on TPUs. Passionate about democratizing Machine Learning.
As often in machine learning, the math is both fun and, in the end, pretty useless 😝.
Let's set the stage: wether playing pong (single-player against a deterministic algorithm) or predicting the next token or sentence in a chain-of-thought LLM, the idea is the same:
Let's set the stage: wether playing pong (single-player against a deterministic algorithm) or predicting the next token or sentence in a chain-of-thought LLM, the idea is the same:
March 11, 2025 at 1:39 AM
As often in machine learning, the math is both fun and, in the end, pretty useless 😝.
Let's set the stage: wether playing pong (single-player against a deterministic algorithm) or predicting the next token or sentence in a chain-of-thought LLM, the idea is the same:
Let's set the stage: wether playing pong (single-player against a deterministic algorithm) or predicting the next token or sentence in a chain-of-thought LLM, the idea is the same: