HarmAug: Effective Data Augmentation for Knowledge Distillation of Safety Guard Models Paper ā¢ 2410.01524 ā¢ Published Oct 2, 2024 ā¢ 3
view post Post 1193 šØš„ New Release Alert! š„šØIntroducing the 435M model that outperforms Llama-Guard-3-8B while slashing 75% of the computation cost! š»š„š Check it out: hbseong/HarmAug-Guard (Yes, INFERENCE CODE INCLUDED! š”)More details in our paper: https://arxiv.org/abs/2410.01524 š#HarmAug #LLM # Safety #EfficiencyBoost #Research #AI #MachineLearning š 5 5 ā¤ļø 4 4 + Reply
Self-Supervised Dataset Distillation for Transfer Learning Paper ā¢ 2310.06511 ā¢ Published Oct 10, 2023
DiffusionNAG: Predictor-guided Neural Architecture Generation with Diffusion Models Paper ā¢ 2305.16943 ā¢ Published May 26, 2023
Knowledge-Augmented Reasoning Distillation for Small Language Models in Knowledge-Intensive Tasks Paper ā¢ 2305.18395 ā¢ Published May 28, 2023 ā¢ 1
Contrastive Learning with Adversarial Perturbations for Conditional Text Generation Paper ā¢ 2012.07280 ā¢ Published Dec 14, 2020
Effective and Efficient Conversation Retrieval for Dialogue State Tracking with Implicit Text Summaries Paper ā¢ 2402.13043 ā¢ Published Feb 20, 2024 ā¢ 2
Self-Distillation for Further Pre-training of Transformers Paper ā¢ 2210.02871 ā¢ Published Sep 30, 2022 ā¢ 1
HarmAug: Effective Data Augmentation for Knowledge Distillation of Safety Guard Models Paper ā¢ 2410.01524 ā¢ Published Oct 2, 2024 ā¢ 3
Scalable Set Encoding with Universal Mini-Batch Consistency and Unbiased Full Set Gradient Approximation Paper ā¢ 2208.12401 ā¢ Published Aug 26, 2022