List of supported LLMs for generation

See detector_benchmark/generation/gen_loader.py

Supported LLMs

• “qwen2_0_5B”: Qwen 2 0.5B non-chat tuned (Qwen/Qwen2-0.5B)

• “qwen2_chat_0_5B”: Qwen 2 0.5B chat tuned (Qwen/Qwen2-0.5B-Instruct)

• “qwen2_chat_7B”: Qwen 2 7B chat tuned (Qwen/Qwen2-7B-Instruct)

• “zephyr”: Zephyr 7B Beta version (HuggingFaceH4/zephyr-7b-beta)

• “llama3_instruct_3”: Llama 3 8B chat tuned version (meta-llama/Meta-Llama-3-8B-Instruct)

• “llama3_instruct_3.1”: Llama 3.1 8B chat tuned version (meta-llama/Meta-Llama-3.1-8B-Instruct)

• “llama3_3.1”: Llama 3.1 8B non-chat tuned version (meta-llama/Meta-Llama-3.1-8B)