abacusai/Smaug-Llama-3-70B-Instruct-32K

The abacusai/Smaug-Llama-3-70B-Instruct-32K is a 70 billion parameter instruction-tuned causal language model developed by Abacus.AI, built upon Meta's Llama 3 architecture. This version extends the context window to 32K tokens using PoSE and LoRA adapter transfer, enhancing its ability to process longer inputs. It demonstrates strong performance on benchmarks like Arena-Hard, positioning it as a competitive model for complex conversational and reasoning tasks requiring extended context.