22What is instruction tuning, and how does it differ from pretraining and alignment?▼medium★ EssentialOpenAIAnthropicGoogle1 replies◆ premiumInstruction tuning is what turns a raw next-token predictor into something that follows instructions. The signal is placing it correctly in the pretrain to SFT to alignment pipeline and knowing exactly what it does and does not fix.Open full answer →
73How do you curate and filter a supervised fine-tuning (SFT) dataset, and why does a smaller clean set often win?▼hardAnthropicMetaHugging Face1 replies◆ premiumA few thousand carefully chosen examples can beat a million scraped ones. The signal is knowing which filters matter, how you measure example quality, and why diversity beats raw volume.Open full answer →
78Compare distillation recipes for LLMs: hard-label SFT, on-policy logit matching, and rejection sampling.▼hardGoogleMetaHugging Face1 replies◆ premiumDistillation is not one method. The signal is knowing when to match logits versus train on generated text, why on-policy distillation beats off-policy, and how reasoning models are distilled.Open full answer →
