Tool use / function calling

NLSI

Superseded baseline#43 of 55 most-superseded

Cited as a baseline — critiqued by newer work, not yet beaten on a benchmark here

1 papers critique it · 0 beat it on benchmarks

What papers say

Verbatim critique sentences, each from a paper that cites NLSI as a baseline.

They assume preferences are explicitly available (e.g., profiles or past API calls), failing to address the more realistic setting where preferences remain latent and must be inferred from interaction history.
Latent Preference Modeling for Cross-Session Personalized Tool Calling

What to use instead

Recent methods in the same sub-problem, not yet superseded in the knowledge base — arXiv benchmark leaders, not vetted production recommendations.