AFaCTA: Assisting the Annotation of Factual Claim Detection with Reliable LLM Annotators
Paper • 2402.11073 • Published
# Load model directly
from transformers import AutoTokenizer, AutoModelForSequenceClassification
tokenizer = AutoTokenizer.from_pretrained("JingweiNi/roberta-base-afacta")
model = AutoModelForSequenceClassification.from_pretrained("JingweiNi/roberta-base-afacta")Quick Links
RoBERTa-base fine-tuned on PoliClaim_{gold} and PoliClaim_{silver} proposed by AFaCTA paper .
PoliClaim dataset can be found at https://github.com/EdisonNi-hku/AFaCTA
To use it: provide the target sentence and its surrounding two sentences as contexts, where RoBERTa separating token </s> is used to separate sentences
For example: To you, the people of Alabama and the men and women of the Legislature: You are the reason for our progress. </s> This evening, I renew my commitment to you that we will not only continue tackling old problems. </s> We will work together as Alabamians to find new solutions so that our state is the best place to live, work and raise a family for years to come.
- Downloads last month
- 15
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="JingweiNi/roberta-base-afacta")