19 large language models for safety or danger

"For every project that needs guardrails, there's another one where they just get in the way. Some projects demand an LLM that returns the complete, unvarnished truth. For these situations, developers are creating unfettered LLMs that can interact without reservation. Some of these solutions are based on entirely new models while others remove or reduce the guardrails built into popular open source LLMs."

"Some scientists are building LLMs that can act as guardrails. These new models are specially trained to recognize when an LLM is potentially going off the rails. If they don't like how an interaction is going, they have the power to stop it."

"The models in this category emphasize the many dimensions of AI safety. Whether you are looking for an LLM built for sensitive topics, one with a strong ethical compass, or a model capable of recognizing hidden exploits in seemingly innocent prompts, the heavily guarded models in this list could have you covered."

Large language models present a fundamental tension: some projects require robust guardrails to prevent harmful outputs, while others demand unrestricted, unvarnished responses. Scientists are developing specialized LLMs trained to recognize when other LLMs go off-track and stop problematic interactions. Simultaneously, developers create unfettered models without safety restrictions for projects requiring complete transparency. Solutions range from entirely new models to modifications of existing open-source LLMs. The landscape includes heavily guarded models emphasizing AI safety across multiple dimensions, designed for sensitive topics and ethical considerations, alongside models that strip away guardrails entirely. This represents the current state-of-the-art in balancing safety requirements with operational freedom.

Read at InfoWorld