AdarshSingh7647/Std-SFT-Passage-8B
AdarshSingh7647/Std-SFT-Passage-8B is an 8 billion parameter Qwen3-based passage reranker model developed by AdarshSingh7647. It is specifically trained using the Std-SFT method, which optimizes the loss on the teacher's reasoning and the final answer. This model functions as a pointwise reranker, determining passage relevance to a query by generating a 'true' or 'false' reasoning, and is particularly effective for information retrieval tasks.
Loading preview...
Std-SFT-Passage-8B Overview
AdarshSingh7647/Std-SFT-Passage-8B is an 8 billion parameter passage reranker built upon the Qwen3 architecture. This model is distinguished by its training methodology, Std-SFT (Standard Supervised Fine-Tuning), where the optimization loss is applied to both the teacher's reasoning process and the ultimate answer. This approach aims to enhance the model's ability to accurately assess the relevance of passages to a given query.
Key Capabilities
- Passage Reranking: Efficiently reorders a list of passages based on their relevance to a specific query.
- Reasoning-based Relevance: Determines relevance by generating an internal reasoning process, concluding with either
trueorfalse. - Pointwise Reranking: Operates as a pointwise reranker, similar to models like Rank1, evaluating each passage independently against the query.
- Performance: Demonstrates competitive performance in reranking tasks, achieving a BRIGHT nDCG@10 score of 31.5, as shown in comparative benchmarks against other Qwen3 variants and GLM-Z1-9B.
Good for
- Information Retrieval Systems: Ideal for improving the quality of search results by re-ranking retrieved documents or passages.
- Question Answering Systems: Can be integrated to select the most relevant passages for generating answers.
- Semantic Search: Enhancing the precision of search queries by identifying semantically relevant content.
- Research and Knowledge Discovery: Assisting users in navigating large corpuses of text by prioritizing pertinent information.