necrashter/interchangeable-token-embeddings-datasets
收藏Hugging Face2026-01-31 更新2026-03-29 收录
下载链接：
https://hf-mirror.com/datasets/necrashter/interchangeable-token-embeddings-datasets
下载链接
链接失效反馈官方服务：
资源简介：
---
license: apache-2.0
---
# Interchangeable Token Embeddings - Datasets
This repository contains the datasets for the ICML 2025 paper, [Interchangeable Token Embeddings for Extendable Vocabulary and Alpha-Equivalence](https://arxiv.org/abs/2410.17161).
Links:
1. [arXiv](https://arxiv.org/abs/2410.17161)
2. [Project Page](https://necrashter.github.io/interchangeable-token-embeddings/)
3. [Code](https://github.com/necrashter/interchangeable-token-embeddings)
4. [Models](https://huggingface.co/necrashter/interchangeable-token-embeddings)
## Structure
* `data`: Datasets for the LTL task
* `na-*`: Datasets from the DeepLTL paper
* `ltl-35-10ap`: Dataset for the full vocabulary baseline
* `ltl-35-perturbed`: The training dataset used in perturbation experiments
* `eval2d-10ap.pkl`: The test dataset used in the heatmap evaluations
* `data-prop`: Datasets for the propositional logic task
* Same internal structure
The datasets for the copy task are omitted since those are just random strings and unnecessarily large, but I can upload them upon request.
提供机构：
necrashter
