RichardErkhov/GraySwanAI_-_Llama-3-8B-Instruct-RR-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

Llama-3-8B-Instruct-RR - GGUF

• Model creator: https://huggingface.co/GraySwanAI/
• Original model: https://huggingface.co/GraySwanAI/Llama-3-8B-Instruct-RR/

Name	Quant method	Size
Llama-3-8B-Instruct-RR.Q2_K.gguf	Q2_K	2.96GB
Llama-3-8B-Instruct-RR.IQ3_XS.gguf	IQ3_XS	3.28GB
Llama-3-8B-Instruct-RR.IQ3_S.gguf	IQ3_S	3.43GB
Llama-3-8B-Instruct-RR.Q3_K_S.gguf	Q3_K_S	3.41GB
Llama-3-8B-Instruct-RR.IQ3_M.gguf	IQ3_M	3.52GB
Llama-3-8B-Instruct-RR.Q3_K.gguf	Q3_K	3.74GB
Llama-3-8B-Instruct-RR.Q3_K_M.gguf	Q3_K_M	3.74GB
Llama-3-8B-Instruct-RR.Q3_K_L.gguf	Q3_K_L	4.03GB
Llama-3-8B-Instruct-RR.IQ4_XS.gguf	IQ4_XS	4.18GB
Llama-3-8B-Instruct-RR.Q4_0.gguf	Q4_0	4.34GB
Llama-3-8B-Instruct-RR.IQ4_NL.gguf	IQ4_NL	4.38GB
Llama-3-8B-Instruct-RR.Q4_K_S.gguf	Q4_K_S	4.37GB
Llama-3-8B-Instruct-RR.Q4_K.gguf	Q4_K	4.58GB
Llama-3-8B-Instruct-RR.Q4_K_M.gguf	Q4_K_M	4.58GB
Llama-3-8B-Instruct-RR.Q4_1.gguf	Q4_1	4.78GB
Llama-3-8B-Instruct-RR.Q5_0.gguf	Q5_0	5.21GB
Llama-3-8B-Instruct-RR.Q5_K_S.gguf	Q5_K_S	5.21GB
Llama-3-8B-Instruct-RR.Q5_K.gguf	Q5_K	5.34GB
Llama-3-8B-Instruct-RR.Q5_K_M.gguf	Q5_K_M	5.34GB
Llama-3-8B-Instruct-RR.Q5_1.gguf	Q5_1	5.65GB
Llama-3-8B-Instruct-RR.Q6_K.gguf	Q6_K	6.14GB
Llama-3-8B-Instruct-RR.Q8_0.gguf	Q8_0	7.95GB

Original model description:

Model Details

Llama-3-8B-Instruct-RR is a Llama-3 model with circuit breakers inserted using Representation Rerouting (RR).

Circuit Breaking is a new approach inspired by representation engineering, designed to prevent AI systems from generating harmful content by directly altering harmful model representations, with minimal capability degradation. For more information, please check out our paper.