arXiv cs.CR·4d agoSeal, Then Sample: Sampled Layerwise Proofs for Verifiable LLM Inference from GPT-2 to 70B#verifiable-inference#sampled-layerwise-proofs#llama-2AI research
Hugging Face daily papers·9d agoThe Functionalizer: Lossless Functional Decomposition for Subword Tokenization#gpt-2#pretokenizer#subword-tokenization
arXiv cs.AI / cs.LG / cs.CL·11d agoRight Tool, Right Job: Native-Language Evaluation, Tokenizer Sensitivity, and Methodological Findings from a French-Only BabyLM#babylm#cross-lingual#evaluationAI research
Hacker News · AI·13d agoDue to concerns about malicious applications, GPT2 will not be released (2019)#gpt-2#language-model#misuse 12 min
