A 109M-parameter bi-encoder with token-level attention, ConTeXT-match, achieves state-of-the-art skill extraction on most ESCO benchmarks at a fraction of the cost of LLM pipelines.
JobBERT : understanding job titles through skills
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
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Efficient Text Encoders for Labor Market Analysis
A 109M-parameter bi-encoder with token-level attention, ConTeXT-match, achieves state-of-the-art skill extraction on most ESCO benchmarks at a fraction of the cost of LLM pipelines.