List your company
ResearchFormerly WizWand

Argument Component Detection & Classification on ElecDeb (test)

63Macro F1RooseBERT
5.820.6535.550.35Aug 5, 2025
Updated 8mo ago

Evaluation results

MethodLinks
RooseBERTCasing=cased, Pre-training Strategy=CONT2025.0863
RooseBERTCasing=cased, Pre-training Strategy=SCR2025.0863
RooseBERTCasing=uncased, Pre-training Strategy=SCR2025.0863
RooseBERTCasing=cased, Pre-training Strategy=SCR2025.0863
ConfliBERTCasing=cased, Pre-training Strategy=CONT2025.0862
ConfliBERTCasing=uncased, Pre-training Strategy=CONT2025.0862
ConfliBERTCasing=cased, Pre-training Strategy=SCR2025.0862
ConfliBERTCasing=uncased, Pre-training Strategy=SCR2025.0862
RooseBERTCasing=uncased, Pre-training Strategy=CONT2025.0862
RooseBERTCasing=uncased, Pre-training Strategy=SCR2025.0862
BERTCasing=cased, Pre-training Strategy=Standard2025.0861
BERTCasing=uncased, Pre-training Strategy=Standard2025.0861
ConfliBERTCasing=cased, Pre-training Strategy=CONT2025.0861
ConfliBERTCasing=uncased, Pre-training Strategy=CONT2025.0861
RooseBERTCasing=cased, Pre-training Strategy=CONT2025.0861
RooseBERTCasing=uncased, Pre-training Strategy=CONT2025.0861
PoliBERTweetCasing=cased, Pre-training Strategy=CONT2025.0860
BERTCasing=cased, Pre-training Strategy=Standard2025.0858
BERTCasing=uncased, Pre-training Strategy=Standard2025.0858
ConfliBERTCasing=cased, Pre-training Strategy=SCR2025.0858
PoliBERTweetCasing=cased, Pre-training Strategy=CONT2025.0854
ConfliBERTCasing=uncased, Pre-training Strategy=SCR2025.0835
gpt-4.1-nanoEvaluation Protocol=one-shot2025.0828
gpt-4.1-nanoEvaluation Protocol=3-shot2025.0826
Mistral-7B-Instruct-v0.3Evaluation Protocol=one-shot2025.0822
Mistral-7B-Instruct-v0.3Evaluation Protocol=3-shot2025.0822
Llama-3.1-8B-InstructEvaluation Protocol=3-shot2025.0822
gemma-3-4b-itEvaluation Protocol=one-shot2025.0821
gemma-3-4b-itEvaluation Protocol=3-shot2025.0821
Llama-3.1-8B-InstructEvaluation Protocol=one-shot2025.0816
gemma-3-4b-itEvaluation Protocol=fine-tuning2025.088
Mistral-7B-Instruct-v0.3Evaluation Protocol=fine-tuning2025.088
Llama-3.1-8B-InstructEvaluation Protocol=fine-tuning2025.088