SuperNI

The 'SuperNI' benchmark is used to evaluate the performance of language models on general natural language processing tasks while ensuring that their capabilities in computational argumentation remain stable.

Papers using SuperNI (3)

ArgInstruct: Specialized Instruction Fine-Tuning for Computational Argumentation2025CrossICL: Cross-Task In-Context Learning via Unsupervised Demonstration Transfer2025In-Context Transfer Learning: Demonstration Synthesis by Transferring Similar Tasks2024