6 papers
Prism-Reranker: Beyond Relevance Scoring -- Jointly Producing Contributions and Evidence for Agentic Retrieval
Dun Zhang
Modern retrieval pipelines increasingly serve downstream consumers like retrieval-augmented generation (RAG) and autonomous agents that need more than a scalar relevance score. A r…
PosIR: Position-Aware Heterogeneous Information Retrieval Benchmark
Ziyang Zeng, Dun Zhang, Yu Yan +4
In real-world documents, the information relevant to a user query may reside anywhere from the beginning to the end. This makes position bias -- a systematic tendency of retrieval…
Jasper-Token-Compression-600M Technical Report
Dun Zhang, Ziyang Zeng, Yudong Zhou +1
This technical report presents the training methodology and evaluation results of the open-source Jasper-Token-Compression-600M model, released in November 2025. Building on previo…
An Empirical Study of Position Bias in Modern Information Retrieval
Ziyang Zeng, Dun Zhang, Jiacheng Li +3
This study investigates the position bias in information retrieval, where models tend to overemphasize content at the beginning of passages while neglecting semantically relevant i…
Dewey Long Context Embedding Model: A Technical Report
Dun Zhang, Panxiang Zou, Yudong Zhou
This technical report presents the training methodology and evaluation results of the open-source dewey_en_beta embedding model. The increasing demand for retrieval-augmented gener…
Jasper and Stella: distillation of SOTA embedding models
Dun Zhang, Jiacheng Li, Ziyang Zeng +1
A crucial component in many deep learning applications, such as Frequently Asked Questions (FAQ) and Retrieval-Augmented Generation (RAG), is dense retrieval. In this process, embe…