Reasoning
Strengthening LLM reasoning by fine-tuning, reinforcement learning and inference scaling for diverse solutions and robust inference.
Text Reasoning and Analysis (TeRA) Lab is a research group at the Visual Computing Center, College of Computer Science and Software Engineering, Shenzhen University. TeRA is dedicated to research in natural language processing and large language models, exploring text-oriented AI technologies such as reasoning, representation, analysis, and agents.
We recruit several incoming Master&PhD students and also undergraduate students as interns each year. See page join us for details.
Strengthening LLM reasoning by fine-tuning, reinforcement learning and inference scaling for diverse solutions and robust inference.
Building compact, effective representations and keeping them interpretable across tasks, such as retrieval and semantic similarity.
Locating, steering, and refining AI behavior using interpretable units and causal methods for mechanism-guided tuning.
Designing memory, tool use, and interactions for long-context AI agents that act reliably across real-world tasks.
L6-713, Zhizhen Building, Canghai Campus of Shenzhen University
Map tiles © OpenStreetMap contributors
