Open Domain Inference With Semantic Api Retrieval

via “open-domain inference with semantic api retrieval”

Framework for training LLM agents on 16K+ real APIs.

Unique: Learns a dedicated API retriever component that ranks 16,464 APIs by semantic relevance to queries, enabling open-domain tool use without explicit API specification, rather than requiring users to specify tools upfront or using simple keyword matching.

vs others: Semantic API retrieval outperforms keyword-based tool selection (e.g., BM25) on diverse queries, and enables discovery of APIs with non-obvious names or descriptions that keyword matching would miss.

via “cross-domain-semantic-transfer”

sentence-similarity model by undefined. 23,35,18,673 downloads.

Unique: Trained via multi-task learning on 8+ heterogeneous datasets (S2ORC papers, MS MARCO web search, StackExchange Q&A, Yahoo Answers, CodeSearchNet, SearchQA, ELI5) rather than single-domain optimization, creating a 'semantic commons' that generalizes across task boundaries at the cost of domain-specific peak performance

vs others: Better zero-shot transfer to unseen domains than domain-specific embeddings (e.g., SciBERT for papers only), though 5-15% lower performance than fine-tuned models on specialized tasks; more practical for multi-domain applications than maintaining separate embedding models