HyperAI초신경

홈 뉴스 최신 연구 논문 튜토리얼 데이터셋 백과사전 SOTA LLM 모델 GPU 랭킹 컨퍼런스

한국어

HyperAI초신경

Few Shot Text Classification On Raft

평가 지표

Over

ADE

Avg

B77

NIS

OSE

SOT

SRI

TAI

TC

TEH

ToS

평가 결과

이 벤치마크에서 각 모델의 성능 결과

모델 이름	Over	ADE	Avg	B77	NIS	OSE	SOT	SRI	TAI	TC	TEH	ToS	Paper Title	Repository
GPT-3 zero-shot	0.378	0.163	0.292	0.000	0.572	0.323	0.628	0.027	0.362	0.290	0.303	0.164	RAFT: A Real-World Few-Shot Text Classification Benchmark
T-Few	0.95	0.804	0.758	0.695	0.833	0.676	0.915	0.508	0.736	0.879	0.586	0.75	Few-Shot Parameter-Efficient Fine-Tuning is Better and Cheaper than In-Context Learning
Plurality-class	0.337	0.446	0.331	0.000	0.353	0.164	0.271	0.493	0.344	0.391	0.366	0.471	RAFT: A Real-World Few-Shot Text Classification Benchmark
GPT-2	0.498	0.600	0.458	0.121	0.561	0.245	0.380	0.492	0.612	0.723	0.311	0.498	RAFT: A Real-World Few-Shot Text Classification Benchmark
AdaBoost	0.838	0.543	0.514	0.023	0.626	0.475	0.455	0.506	0.556	0.625	0.443	0.560	RAFT: A Real-World Few-Shot Text Classification Benchmark
BART MNLI zero-shot	0.462	0.234	0.382	0.332	0.615	0.360	0.644	0.026	0.469	0.400	0.543	0.122	RAFT: A Real-World Few-Shot Text Classification Benchmark
GPT-3	0.937	0.686	0.627	0.299	0.679	0.431	0.769	0.516	0.656	0.821	0.526	0.574	RAFT: A Real-World Few-Shot Text Classification Benchmark
GPT-Neo	0.681	0.452	0.481	0.149	0.408	0.343	0.406	0.493	0.605	0.636	0.554	0.565	RAFT: A Real-World Few-Shot Text Classification Benchmark
Human (crowdsourced)	0.917	0.830	0.735	0.607	0.857	0.646	0.908	0.468	0.609	0.897	0.722	0.627	RAFT: A Real-World Few-Shot Text Classification Benchmark

0 of 9 row(s) selected.