its_just_andy@alien.topB to LocalLLaMA@poweruser.forumEnglish · 2 years agoUsing simple tree-search techniques for LLM token sampling can give better resultsandys.pageexternal-linkmessage-square1linkfedilinkarrow-up11arrow-down10
arrow-up11arrow-down1external-linkUsing simple tree-search techniques for LLM token sampling can give better resultsandys.pageits_just_andy@alien.topB to LocalLLaMA@poweruser.forumEnglish · 2 years agomessage-square1linkfedilink
minus-squarekilust@alien.topBlinkfedilinkEnglisharrow-up1·2 years agoThat’s really interesting! Maybe an experiment with the viterbi algorithm with the goal to maximize similarity with the context (or another score) could also improve results 🤔
That’s really interesting! Maybe an experiment with the viterbi algorithm with the goal to maximize similarity with the context (or another score) could also improve results 🤔