Text Classification
Transformers
Safetensors
English
bert
prompt-compression
dependency-detection
referential-dangling
research
text-embeddings-inference
Instructions to use JusperLee/referential-dangling-detector with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use JusperLee/referential-dangling-detector with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="JusperLee/referential-dangling-detector")# Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("JusperLee/referential-dangling-detector") model = AutoModelForSequenceClassification.from_pretrained("JusperLee/referential-dangling-detector", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Add paper author information
Browse files
README.md
CHANGED
|
@@ -24,6 +24,17 @@ The model scores whether a candidate sentence supplies a necessary dependency
|
|
| 24 |
for a retained sentence, conditioned on the question. It is used by the
|
| 25 |
repository's automatic context-restoration experiments.
|
| 26 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 27 |
## Model details
|
| 28 |
|
| 29 |
- **Architecture:** BERT sequence classifier with two labels
|
|
@@ -83,6 +94,7 @@ languages, or substantially different compression settings without evaluation.
|
|
| 83 |
```bibtex
|
| 84 |
@misc{referentialdangling,
|
| 85 |
title={Relevant but Incomplete: Referential Dangling as a Paradigm-Level Failure Mode in Hard Prompt Compression},
|
|
|
|
| 86 |
note={Research code and model release}
|
| 87 |
}
|
| 88 |
```
|
|
|
|
| 24 |
for a retained sentence, conditioned on the question. It is used by the
|
| 25 |
repository's automatic context-restoration experiments.
|
| 26 |
|
| 27 |
+
## Authors
|
| 28 |
+
|
| 29 |
+
Zhengpei Hu<sup>1,∗</sup>, Kai Li<sup>2,∗</sup>, Dapeng Fu<sup>3</sup>,
|
| 30 |
+
Xuechao Zou<sup>2</sup>, Yuanhao Tang<sup>1</sup>, Yue Li<sup>1</sup>,
|
| 31 |
+
Tengfei Cao<sup>1</sup>, and Jianqiang Huang<sup>1,†</sup>
|
| 32 |
+
|
| 33 |
+
- <sup>1</sup>School of Computer Technology and Application, Qinghai University
|
| 34 |
+
- <sup>2</sup>Tsinghua University
|
| 35 |
+
- <sup>3</sup>Ant Group Security and Intelligence Laboratory (SIL)
|
| 36 |
+
- <sup>∗</sup>Equal contribution; <sup>†</sup>corresponding author
|
| 37 |
+
|
| 38 |
## Model details
|
| 39 |
|
| 40 |
- **Architecture:** BERT sequence classifier with two labels
|
|
|
|
| 94 |
```bibtex
|
| 95 |
@misc{referentialdangling,
|
| 96 |
title={Relevant but Incomplete: Referential Dangling as a Paradigm-Level Failure Mode in Hard Prompt Compression},
|
| 97 |
+
author={Zhengpei Hu and Kai Li and Dapeng Fu and Xuechao Zou and Yuanhao Tang and Yue Li and Tengfei Cao and Jianqiang Huang},
|
| 98 |
note={Research code and model release}
|
| 99 |
}
|
| 100 |
```
|