Datasets:

Modalities:
Text
Formats:
parquet
ArXiv:
Libraries:
Datasets
pandas
License:
File size: 1,292 Bytes
463cc34
 
c9d15d9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
463cc34
fd7d1ef
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
---
license: mit
dataset_info:
  features:
  - name: 1 Round Prompt
    dtype: string
  - name: 2 Round Prompt
    dtype: string
  - name: 3 Round Prompt
    dtype: string
  - name: 4 Round Prompt
    dtype: string
  - name: 5 Round Prompt
    dtype: string
  - name: label
    dtype: string
  splits:
  - name: GoEmotion
    num_bytes: 22269392
    num_examples: 500
  - name: BANKING77
    num_bytes: 63903578
    num_examples: 500
  - name: FewNERD
    num_bytes: 117612568
    num_examples: 500
  - name: TacRED
    num_bytes: 35790846
    num_examples: 500
  - name: Discovery
    num_bytes: 353082806
    num_examples: 500
  - name: DialogRE
    num_bytes: 35512103
    num_examples: 118
  download_size: 281420232
  dataset_size: 628171293
configs:
- config_name: default
  data_files:
  - split: GoEmotion
    path: data/GoEmotion-*
  - split: BANKING77
    path: data/BANKING77-*
  - split: FewNERD
    path: data/FewNERD-*
  - split: TacRED
    path: data/TacRED-*
  - split: Discovery
    path: data/Discovery-*
  - split: DialogRE
    path: data/DialogRE-*
---
This is the benchmark we adopt in [Long-context LLMs Struggle with Long In-context Learning](https://arxiv.org/abs/2404.02060). Check out our leaderboard at https://huggingface.co/spaces/TIGER-Lab/LongICL-Leaderboard.