Apart from the fact it's focused on logical fallacies, this is reminiscent of AWS Bedrock Automated Reasoning, which also appears to involve some kind of LLM-gu

Apart from the fact it's focused on logical fallacies, this is reminiscent of AWS Bedrock Automated Reasoning, which also appears to involve some kind of LLM-guided translation of natural language into logical rules ... which are then used to validate the output of the LLM application https://aws.amazon.com/blogs/aws/prevent-factual-errors-from-llm-hallucinations-with-mathematically-sound-automated-reasoning-checks-preview/ https://aws.amazon.com/blogs/aws/prevent-factual-errors-from...