On six public mental health text datasets, small language models reach macro F1 within 0.02 of large LLM baselines on binary tasks, but the comparison relies on previously published scores.
A taxonomy of ethical tensions in inferring mental health states from social media,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2025 1verdicts
REJECT 1roles
background 1polarities
support 1representative citing papers
citing papers explorer
-
Beyond Scale: Small Language Models are Comparable to GPT-4 in Mental Health Understanding
On six public mental health text datasets, small language models reach macro F1 within 0.02 of large LLM baselines on binary tasks, but the comparison relies on previously published scores.