Assessing Open-Source Large Language Models on Argumentation Mining Subtasks
This work provides insights into the performance of open-source LLMs for computational argumentation, which is incremental as it benchmarks existing models on new data without introducing novel methods.
The paper assessed the capability of four open-source large language models (Mistral 7B, Mixtral8x7B, LlamA2 7B, LlamA3 8B) on argumentation mining subtasks like argumentative discourse units and relation classification across three corpora in zero-shot and few-shot scenarios, finding that these models can perform competitively but with varying results depending on the task and dataset.
We explore the capability of four open-sourcelarge language models (LLMs) in argumentation mining (AM). We conduct experiments on three different corpora; persuasive essays(PE), argumentative microtexts (AMT) Part 1 and Part 2, based on two argumentation mining sub-tasks: (i) argumentative discourse units classifications (ADUC), and (ii) argumentative relation classification (ARC). This work aims to assess the argumentation capability of open-source LLMs, including Mistral 7B, Mixtral8x7B, LlamA2 7B and LlamA3 8B in both, zero-shot and few-shot scenarios. Our analysis contributes to further assessing computational argumentation with open-source LLMs in future research efforts.