Llama 3.2 mlp pruned Created by pruning the MLP (feedforward) layers, reducing the size of Llama models while improving their performance. oopere/pruned20-llama-1b Text Generation • Updated 14 days ago • 52 oopere/pruned40-llama-1b Text Generation • Updated 14 days ago • 56 oopere/pruned60-llama-1b Text Generation • Updated 24 days ago • 39
Interesting LLM Papers Soft-prompt Tuning for Large Language Models to Evaluate Bias Paper • 2306.04735 • Published Jun 7, 2023 • 1
Soft-prompt Tuning for Large Language Models to Evaluate Bias Paper • 2306.04735 • Published Jun 7, 2023 • 1