File size: 1,742 Bytes
18fcf61 cb7e01c |
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 56 57 |
---
dataset_info:
features:
- name: hyponym
dtype: string
- name: hypernym
dtype: string
- name: definition
dtype: string
splits:
- name: train
num_bytes: 4974231
num_examples: 44772
- name: test
num_bytes: 5422
num_examples: 49
download_size: 3485218
dataset_size: 4979653
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
---
# Dataset card for WordNet-TaxoLLaMA
TaxoLLaMA is a model capable of solving Lexical Semantics task with SoTA metrics.
The model was fine-tuned on instructive dataset WordNet-TaxoLLaMA. It consists of hypernym-hyponym pairs sampled from WordNet 3.0. As well, it contains definitions, that were used during training to help model disambiguate senses.
## Input Format
The TaxoLLaMA model was trained to use the following format :
```
<s>[INST] <<SYS>> You are a helpfull assistant. List all the possible words divided with a coma. Your answer should not include anything except the words divided by a coma<</SYS>>
hyponym: tiger (large feline of forests in most of Asia having a tawny coat with black stripes)| hypernyms: [/INST]
```
We recommend you to follow this format, however you are free to change it to suite your task!
## Citation
If you find TaxoLLaMA or WordNet-TaxoLLaMA is useful in your work, please cite it with:
```
@misc{moskvoretskii2024taxollama,
title={TaxoLLaMA: WordNet-based Model for Solving Multiple Lexical Sematic Tasks},
author={Viktor Moskvoretskii and Ekaterina Neminova and Alina Lobanova and Alexander Panchenko and Irina Nikishina},
year={2024},
eprint={2403.09207},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
|