arxiv:2308.09138
Domenic Rosati
domenicrosati
AI & ML interests
Natural Language Processing on Scientific Texts, Theory grounded NLP/NLU, Representation Learning, Safety, Factuality, Truth Conditions, Unverstanding and Meaning
Organizations
Papers
2
models
197
domenicrosati/repnoise_beta0.001_attacked_2
Text Generation
•
Updated
•
1
domenicrosati/repnoise_beta0.001_2
Feature Extraction
•
Updated
•
1
domenicrosati/repnoise_0.001beta_attacked_3e-4
Text Generation
•
Updated
•
13
domenicrosati/repnoise_0.001_beta
Text Generation
•
Updated
•
180
domenicrosati/repremove
Text Generation
•
Updated
•
19
domenicrosati/representationremoval
Updated
domenicrosati/security_vectors_meta-llama_Llama-2-7b-chat-hf_8e-5_10k
Updated
domenicrosati/security_vectors_meta-llama_Llama-2-7b-chat-hf_3e-5_1k
Updated
domenicrosati/freeze_layers_ten_twenty_meta-llama_Llama-2-7b-chat-hf_minimality-mmd_defence_steps_10000
Text Generation
•
Updated
domenicrosati/freeze_layers_lm_head_4_meta-llama_Llama-2-7b-chat-hf_adversarial_loss_defence_steps_10000
Text Generation
•
Updated