Maybe it is in my subfield (linguistics/psychology), but it gets everything wrong. It fails to identify most supporting and contrasting citation. When it classifies one as such, is oftentimes incorrect (e.g. the citation is not actually supportive) or misleading (e.g. a self-citation of previous research). Scite would be a really cool instrument; but at its current state, it does not seem to work for entire, massive fields of studies.
scite allows users to see how a publication has been cited, providing the citation context and classification
It helps researchers better discover and evaluate scientific articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence
The platform is trusted by leading universities, publishers, and corporations worldwide
It offers a Large Language Model (LLM) Experience for Researchers, providing assistance in search strategies, building reference lists, writing marketing and blog posts, and more
The Assistant by scite is built with observability in mind to help users make more informed decisions about AI-generated content