analiza aiqbench
This study benchmarks instruction-tuned large language models on identifying Correct Information Units in aphasic speech transcripts without gradient-based training. While few-shot prompting achieves competitive F1 scores comparable to human raters for mild-to-moderate cases, zero-shot methods fail and performance degrades significantly with severe aphasia. The findings suggest LLMs are viable as a high-recall component within human-in-the-loop assessment systems but lack the precision required for fully autonomous clinical use.
Do LLMs Reliably Identify Correct Information Units in Aphasic Discourse?
arXiv:2606.15696v1 Announce Type: cross Abstract: Correct Information Units (CIUs) are central to discourse assessment in aphasia because they quantify communicative informativeness rather than linguistic form alone. However, CIU scoring is time intensive and requires trained raters. This study examined whether instruction-tuned large language models (LLMs) can reliably perform token-level CIU classification from aphasic discourse transcripts. Sixteen picture-description transcripts elicited with the Cat Rescue stimulus were annotated for CIU status according to Nicholas and Brookshire (1993). The sample spanned four severity strata: control, mild, moderate, and severe aphasia. Four publicly available instruction-tuned LLMs were benchmarked under zero-shot and two few-shot prompting conditions across five stratified random seeds. Performance was evaluated against consensus human labels using accuracy, precision, recall, F1, and Cohen's kappa. Zero-shot prompting was insufficient across models. In contrast, few-shot prompting yielded substantial gains and produced competitive performance for three viable models. Mean few-shot F1 scores ranged from 0.776 to 0.817 across Llama-3.1-8B, Qwen2.5-7B, and Mistral-7B, with no significant differences between fixed global and per-chunk local example selection. Phi-3-mini was unstable and did not yield reliable performance. Viable models showed high recall but lower precision, suggesting systematic over-classification of tokens as CIUs. Performance also varied by discourse severity, with the weakest results in more severe aphasia. Few-shot LLM prompting can support automated CIU identification without gradient-based task training, but agreement with human annotation remains insufficient for fully autonomous use. These findings support LLM-based CIU scoring as a promising human-in-the-loop component of discourse assessment systems.
Title:Do LLMs Reliably Identify Correct Information Units in Aphasic Discourse?
Abstract:Correct Information Units (CIUs) are central to discourse assessment in aphasia because they quantify communicative informativeness rather than linguistic form alone. However, CIU scoring is time intensive and requires trained raters. This study examined whether instruction-tuned large language models (LLMs) can reliably perform token-level CIU classification from aphasic discourse transcripts. Sixteen picture-description transcripts elicited with the Cat Rescue stimulus were annotated for CIU status according to Nicholas and Brookshire (1993). The sample spanned four severity strata: control, mild, moderate, and severe aphasia. Four publicly available instruction-tuned LLMs were benchmarked under zero-shot and two few-shot prompting conditions across five stratified random seeds. Performance was evaluated against consensus human labels using accuracy, precision, recall, F1, and Cohen's kappa. Zero-shot prompting was insufficient across models. In contrast, few-shot prompting yielded substantial gains and produced competitive performance for three viable models. Mean few-shot F1 scores ranged from 0.776 to 0.817 across Llama-3.1-8B, Qwen2.5-7B, and Mistral-7B, with no significant differences between fixed global and per-chunk local example selection. Phi-3-mini was unstable and did not yield reliable performance. Viable models showed high recall but lower precision, suggesting systematic over-classification of tokens as CIUs. Performance also varied by discourse severity, with the weakest results in more severe aphasia. Few-shot LLM prompting can support automated CIU identification without gradient-based task training, but agreement with human annotation remains insufficient for fully autonomous use. These findings support LLM-based CIU scoring as a promising human-in-the-loop component of discourse assessment systems.
Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
arXiv:2606.15696 [cs.AI]
(or arXiv:2606.15696v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2606.15696 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 350 cuvinte
Subiecte conexe
When Discourse Pressures Conflict: Information Structure in Vision-Language Model Outputs
arXiv · 25 iul. 2026
Global Crises and National Policies: A Large Scale Analysis of Political Content in German Language Online Media
arXiv · 20 aug. 2026
Fine-grained Analysis of Brain-LLM Alignment through Input Attribution
arXiv · 14 oct. 2026
The persuasive power of large language models does not depend on their perceived national origin
arXiv · 3 aug. 2026