SuperAnnotate
/

roberta-large-llm-content-detector

Text Classification

generated_text_detection

llm_content_detection

Model card Files Files and versions Community

SA-Yur-or commited on May 17

Commit

27dc31d

•

1 Parent(s): 577a013

[doc]: minor fix of readme

Files changed (1) hide show

README.md +2 -3

README.md CHANGED Viewed

@@ -37,7 +37,7 @@ Couple of articles about this problem: [*Problems with Synthetic Data*](https://
 - **Model type:** The custom architecture for binary sequence classification based on pre-trained RoBERTa, with a single output label.
 - **Language(s):** Primarily English.
-- **License:** Apache 2.0
 - **Finetuned from model:** [RoBERTa Large](https://huggingface.co/FacebookAI/roberta-large)
 ### Model Sources
@@ -116,8 +116,7 @@ A custom architecture was chosen for its ability to perform binary classificatio
 ## Performance
-The model was evaluated on a benchmark collected from the same datasets used for training, alongside a closed subset of SuperAnnotate. \
-However, there are no direct intersections of samples between the training data and the benchmark. \
 The benchmark comprises 1k samples, with 200 samples per category. \
 The model's performance is compared with open-source solutions and popular API detectors in the table below:

 - **Model type:** The custom architecture for binary sequence classification based on pre-trained RoBERTa, with a single output label.
 - **Language(s):** Primarily English.
+- **License:** [SAIPL](https://huggingface.co/SuperAnnotate/roberta-large-llm-content-detector/blob/main/LICENSE)
 - **Finetuned from model:** [RoBERTa Large](https://huggingface.co/FacebookAI/roberta-large)
 ### Model Sources
 ## Performance
+The model was evaluated on a benchmark consisting of a holdout subset of training data, alongside a closed subset of SuperAnnotate data. \
 The benchmark comprises 1k samples, with 200 samples per category. \
 The model's performance is compared with open-source solutions and popular API detectors in the table below: