Knowledgator Releases GLiFormer: A 575M-Parameter Encoder That Hits 91.10 F1 on Nested JSON Extraction Without Generating Tokens
575M encoder hits 91.10 F1 on nested JSON extraction—without generating tokens. Knowledgator's GLiFormer closes the gap to GPT-scale models on a specific, high-value task.
Why it matters
A smaller, non-generative encoder model achieves near-frontier performance on structured extraction, suggesting task-specific architectures can compete with large generative models on precision work. Relevant for practitioners choosing between API calls and edge-deployable alternatives.
The key facts
9 to knowGLiFormer: 575M parameters
91.10 F1 on nested JSON extraction
Comparison baseline: GPT-5.6-luna at 91.96 F1
Architecture: encoder-only (no token generation)
Value grounding in source spans (interpretability)
GLiFormer Large: 575M parameters
91.10 F1 on nested JSON extraction benchmark
GPT-5.6-luna: 91.96 F1 (comparative baseline)
Encoder-only architecture (no token generation)
Go to the source
MarkTechPostmarktechpost.com
Publisher excerpt: GLiFormer Large scores 91.10 F1 on nested JSON, near GPT-5.6-luna's 91.96, while grounding every value in source spans. The post Knowledgator Releases GLiFormer: A 575M-Parameter Encoder That Hits 91.10 F1 on Nested JSON Extraction Without Generating Tokens appeared first on MarkTechPost.