metadata
configs:
- config_name: contrastive
data_files:
- split: train
path: contrastive/train-*
- split: validation
path: contrastive/validation-*
- split: test
path: contrastive/test-*
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*
dataset_info:
- config_name: contrastive
features:
- name: 'y'
dtype: string
- name: items_set
dtype: string
- name: prompt
dtype: string
- name: N_items
dtype: int64
splits:
- name: train
num_bytes: 63796428
num_examples: 4400
- name: validation
num_bytes: 39152469
num_examples: 3300
- name: test
num_bytes: 39152509
num_examples: 3300
download_size: 45055778
dataset_size: 142101406
- config_name: default
features:
- name: 'y'
dtype: string
- name: items_set
dtype: string
- name: prompt
dtype: string
- name: N_items
dtype: int64
splits:
- name: train
num_bytes: 64093400
num_examples: 4400
- name: validation
num_bytes: 39375100
num_examples: 3300
- name: test
num_bytes: 39375100
num_examples: 3300
download_size: 45159496
dataset_size: 142843600
Dataset Card for "Missing Item Prediction"
https://arxiv.org/abs/2407.13481
@article{sileo2024attention,
title={Attention Overflow: Language Model Input Blur during Long-Context Missing Items Recommendation},
author={Sileo, Damien},
journal={arXiv preprint arXiv:2407.13481},
year={2024}
}