HejuruUbuyobozi bukurikira
Socratic Prompts ituma AI ikwigisha
Ururimi AI
Ururimi AI
Byibuze-Kuri-Byinshi byihutirwa bikuraho ikibazo gikomeye muburyo bukurikiranye bwibibazo byoroshye, kubikemura kugirango buri gisubizo kigaburira igikurikira.
Umva iki gitaboAI-byavuzwe · ~2 min
Nibyingenzi kuko ireka moderi ikemura ibibazo cyane kuruta ingero zerekanwe.
Least-to-Most prompting, introduced by Zhou and colleagues at Google in 2022, has two stages. First, the model is prompted to decompose a complex question into an ordered list of easier subquestions. Second, it solves those subquestions one at a time, appending each solved answer to the context so later steps can build on earlier ones. This differs from chain-of-thought, which reasons in a single pass without explicit decomposition. The authors evaluated symbolic manipulation, math reasoning, and SCAN, a compositional-generalization benchmark built from natural-language commands. For the specific SCAN setup using GPT-3 code-davinci-002 and 14 exemplars, Least-to-Most reached at least 99% accuracy across the tested splits, including the length split, compared with 16% for chain-of-thought prompting. This is a result for that model, prompt setup, and benchmark, not a general accuracy guarantee.
Gukemura ikibazo cyintambwe nyinshi zijambo ubanza gutondekanya ingano yo kubara, hanyuma ukabara muburyo bukurikirana
Imirimo yindimi zihimbano nko guhindura amabwiriza maremare mubikorwa bikurikiranye kuva ingero ngufi
Gusubiza ikibazo cyubushakashatsi bugoye ukagabanyamo ibibazo bito ibisubizo bihuriza hamwe mubisubizo byanyuma
Kwandika progaramu mu kubora mubikorwa byabafasha byakemuwe icyarimwe, buri kimwe cyakoreshejwe nintambwe nyuma
Byibuze-Kuri-Ibitekerezo byinshi ubu bishimangira abakozi benshi nububiko bwububiko bugabanya intego mubice byateganijwe mbere yo gukora. Tegereza imvange hamwe no gukoresha ibikoresho, aho buri subquestion ishobora gukurura calculatrice, gushakisha, cyangwa guhamagara kode, hamwe no kwihagararaho kubindi bisubizo bikomeye. Ubushakashatsi nabwo burimo gushakisha uburyo bwangirika bwihuza nuburemere bwikibazo, no kubuhuza no kugenzura kugirango subanswer yo hambere itari yo yonona bucece urunigi rwose.
Ibaruwa y'amakuru y'ubuntu
Inkuru eshatu zizewe za AI buri gitondo cy'iminsi y'akazi, zanditse mu rurimi rworoshye. Ubuntu iteka ryose, nta kwamamaza.
Imeri imwe buri munsi y'akazi. Kwiyimura mu buryo bumwe. Ntitwigera tugurisha cyangwa ngo dusangire aderesi yawe.
Isuzume
Ibisubizo ako kanya kuri buri kibazo, n'icyemezo gishobora gusangirwa gifite ID yagenzurwa umaze gutsinda isomo.
Fasha uburezi bwa AI bw'ubuntu. AI Understanding ni umuryango utegamiye kuri leta wa 501(c)(3) — nta kwamamaza, nta kwishyura, iteka ryose. Tanga impano
Byibuze-Kuri-Byinshi byihutirwa bikuraho ikibazo gikomeye muburyo bukurikiranye bwibibazo byoroshye, kubikemura kugirango buri gisubizo kigaburira igikurikira. Nibyingenzi kuko ireka moderi ikemura ibibazo cyane kuruta ingero zerekanwe.
Irabanza ikemure ikibazo mubice byateganijwe byoroshye byoroshye, hanyuma bikemura buri kimwe muburyo bukurikiranye.
Urunigi-rw'ibitekerezo impamvu imwe; byibuze-kuri-byeruye kubora no kuboha iminyururu hamwe.
SCAN tests compositional generalization using commands that map language to action sequences; the Least-to-Most paper evaluates whether prompts transfer to harder command compositions.
Subproblems itunganijwe kuburyo buriwese biterwa nibisubizo byakemuwe mbere, kureka ibisubizo byubaka neza.
Buri subanswer yakemuwe yongewe kumurongo wihuta kuburyo intambwe ikurikiraho ifite ibisubizo bakeneye.
Komeza wige
Abandi bayobozi batoranijwe kuriyi ngingo
HejuruUbuyobozi bukurikira
Socratic Prompts ituma AI ikwigisha
Ururimi AI