instruction tuning
/ɪnˈstrʌkʃən ˌtuːnɪŋ/
fine-tuning models specifically on instruction-following examples
instruction tuning in a sentence
“Instruction tuning transformed the base model into a helpful assistant.”
Origin of instruction tuning
Latin instructio building, arrangement + tuning
Related Words
structured output
model output constrained to follow a defined machine-readable schema, commonly a supported subset of JSON Schema
delimiter
a character or tag that marks where one part of a prompt ends and another begins
prompt template
a reusable prompt skeleton whose slots are filled in at run time
negative prompt
an instruction telling an image model what to leave out
self-consistency
sampling several reasoning paths and taking the majority answer
tree-of-thought
exploring branching reasoning paths and backtracking from dead ends