A5006919400- Microstructure and Mechanical Properties of Steels
- Parallel Computing and Optimization Techniques
- Advanced Data Storage Technologies
- Heat and Mass Transfer in Porous Media
- Evaluation and Performance Assessment
- Layered Double Hydroxides Synthesis and Applications
- Human Motion and Animation
- Acoustic Wave Phenomena Research
- Advanced machining processes and optimization
- Nanomaterials and Printing Technologies
- Video Analysis and Summarization
- Magnesium Alloys: Properties and Applications
- Distributed and Parallel Computing Systems
- Image Retrieval and Classification Techniques
- Caching and Content Delivery
- Adsorption and Cooling Systems
- Advanced Fiber Optic Sensors
- 3D IC and TSV technologies
- Handwritten Text Recognition Techniques
- Metal Alloys Wear and Properties
- Metallurgy and Material Forming
- Heat Transfer and Optimization
- Aluminum Alloys Composites Properties
- Aluminum Alloy Microstructure Properties
- Microwave Engineering and Waveguides
Tsinghua University
1988-2024
Large language models (LLMs) typically undergo instruction tuning to enhance alignment. Recent studies emphasize that quality and diversity of data are more crucial than quantity, highlighting the need select diverse, high-quality subsets reduce training costs. However, how evolve these selected alongside development new remains insufficiently explored. To achieve LLMs' ongoing alignment, we introduce Instruction Bank (InsBank), a continuously updated repository integrates latest valuable...
Evaluating models on large benchmarks is very resource-intensive, especially during the period of rapid model evolution. Existing efficient evaluation methods estimate performance target by testing them only a small and static coreset benchmark, which derived from publicly available results source models. These rely assumption that have high prediction consistency with However, we demonstrate it doesn't generalize well in practice. To alleviate inconsistency issue, present TailoredBench,...
This comprehensive study evaluates the performance of OpenAI's o1-preview large language model across a diverse array complex reasoning tasks, spanning multiple domains, including computer science, mathematics, natural sciences, medicine, linguistics, and social sciences. Through rigorous testing, demonstrated remarkable capabilities, often achieving human-level or superior in areas ranging from coding challenges to scientific processing creative problem-solving. Key findings include: -83.3%...
Indexing structures are widely used in modern data-processing applications to support high-performance queries, and there a variety of recent designs specifically optimized for the newly available persistent memory (PM). The primary focus previous PM indexes is on reducing expensive writes persisting data. However, we find that tree-based indexes, because smaller performance gap between random reads real devices, read-intensive tree traversal phase dominates overall latency. This observation...
Porous copper fabricated by unidirectional solidification of metal-gas eutectic can be used to manufacture a special kind micro-channel heat sink. A three dimensional model is developed investigate the transfer in porous However experimental results performance are far less than simulation results. That mainly because some pores nonpenetrative. When modified taking penetration ratio into account consistent with