HyperAI초신경
홈
뉴스
최신 연구 논문
튜토리얼
데이터셋
백과사전
SOTA
LLM 모델
GPU 랭킹
컨퍼런스
전체 검색
소개
한국어
HyperAI초신경
Toggle sidebar
전체 사이트 검색...
⌘
K
홈
SOTA
Conversational Web Navigation
Conversational Web Navigation On Weblinx
Conversational Web Navigation On Weblinx
평가 지표
Element (IoU)
Intent Match
Overall score
Text (F1)
평가 결과
이 벤치마크에서 각 모델의 성능 결과
Columns
모델 이름
Element (IoU)
Intent Match
Overall score
Text (F1)
Paper Title
Repository
GPT-3.5T (Zero-Shot)
8.62
42.77
8.51
3.45
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
S-LLaMA-1.3B
20.54
83.32
23.73
25.85
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
Pix2Act-1.3B
8.28
81.80
16.88
25.21
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
MindAct-3B
16.50
79.89
20.94
23.16
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
Fuyu-8B
15.70
80.07
19.97
22.30
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
Llama-2-13B
22.82
81.91
25.21
26.60
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
GPT-3.5F
18.64
77.56
21.22
22.39
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
MindAct-780M
13.39
75.87
15.13
13.58
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
Flan-T5-780M
15.36
80.02
17.27
14.05
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
MindAct-250M
12.05
74.25
12.63
7.67
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
Pix2Act-282M
6.20
79.71
12.51
16.40
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
S-LLaMA-2.7B
22.60
84.00
25.02
27.17
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
GPT-4T (Zero-Shot)
10.85
41.66
10.72
6.75
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
Flan-T5-250M
14.86
79.69
14.99
9.21
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
Flan-T5-3B
20.31
81.14
23.77
25.75
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
GPT-4V (Zero-Shot)
10.91
42.36
10.45
6.21
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
Llama-2-7B
22.26
82.64
24.57
26.50
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
0 of 17 row(s) selected.
Previous
Next