Bohan Yang
B.Sc. @ Beijing Normal-Hong Kong Baptist University
Computer Science and Technology

I am currently an undergraduate student pursuing a B.Sc. in Computer Science and Technology at Beijing Normal-Hong Kong Baptist University.

I am actively looking for an MPhil/PhD position for Fall 2027, as well as research collaboration and internship opportunities.

My current research interests include large language models, especially trustworthy and high-efficiency LLMs, and 3D Gaussian Splatting. I am currently a research assistant at the Binjiang Institute of Zhejiang University, working with Dr. Wenpeng Xing and Prof. Meng Han.

I also have a strong background in competitive programming, with experience in NOIP and ICPC/CCPC.


Education
  • Beijing Normal-Hong Kong Baptist University
    B.S. in Computer Science and Technology
    Sep. 2023 - present
  • Hangzhou Xuejun High School
    High School Diploma
    Sep. 2020 - Jul. 2023
Experience
  • Binjiang Institute of Zhejiang University
    Research Assistant with Dr. Wenpeng Xing and Prof. Meng Han
    Jan. 2026 - Jun. 2026
    Trustworthy LLMs, 3D Gaussian Splatting, and vision-language models.
Honors & Awards
  • Bronze Medal (51/149), ICPC Asia-East Regional Contest (Hong Kong)
    2024
  • Second Prize (1278/5261, China), NOIP
    2021
  • Codeforces Master, peak rating 2183
News
Released the project page for TriLens: Per-Layer Logit-Lens Entropy for White-Box Hallucination Detection. Project page / Paper
May 26, 2026
Our paper QuadBox: Accelerating 3D Gaussian Splatting with Geometry-Aware Boxes was accepted to ICIP 2026.
May 01, 2026
Awarded a Bronze Medal (51st out of 149 teams) at the ICPC Asia-East Regional Contest in Hong Kong.
Dec 22, 2024
Selected Publications (view all ) * Equal contribution. # Corresponding author.
TriLens: Per-Layer Logit-Lens Entropy for White-Box Hallucination Detection

Bohan Yang, Yijun Gong, Zhi Zhang, Ge Zhang, Wenpeng Xing#, Meng Han#

Under review First Author

TriLens introduces per-layer logit-lens entropy as a white-box signal for hallucination detection in large language models.

TriLens: Per-Layer Logit-Lens Entropy for White-Box Hallucination Detection

Bohan Yang, Yijun Gong, Zhi Zhang, Ge Zhang, Wenpeng Xing#, Meng Han#

Under review First Author

TriLens introduces per-layer logit-lens entropy as a white-box signal for hallucination detection in large language models.

QuadBox: Accelerating 3D Gaussian Splatting with Geometry-Aware Boxes

Xinze Li, Bohan Yang, Pengxu Chen, Yiyuan Wang, Hongcheng Luo, Wentao Cheng#, Weifeng Su#

IEEE International Conference on Image Processing (ICIP) 2026 Published

QuadBox accelerates 3D Gaussian Splatting using geometry-aware boxes for more efficient scene representation and rendering.

QuadBox: Accelerating 3D Gaussian Splatting with Geometry-Aware Boxes

Xinze Li, Bohan Yang, Pengxu Chen, Yiyuan Wang, Hongcheng Luo, Wentao Cheng#, Weifeng Su#

IEEE International Conference on Image Processing (ICIP) 2026 Published

QuadBox accelerates 3D Gaussian Splatting using geometry-aware boxes for more efficient scene representation and rendering.

Latent Fusion Jailbreak: Blending Harmful and Harmless Representations to Elicit Unsafe LLM Outputs

Wenpeng Xing, Bohan Yang, Zaifeng Yang, Changting Lin, Meng Han#

Under review Student First Author

Latent Fusion Jailbreak studies how blending harmful and harmless representations can elicit unsafe outputs from large language models.

Latent Fusion Jailbreak: Blending Harmful and Harmless Representations to Elicit Unsafe LLM Outputs

Wenpeng Xing, Bohan Yang, Zaifeng Yang, Changting Lin, Meng Han#

Under review Student First Author

Latent Fusion Jailbreak studies how blending harmful and harmless representations can elicit unsafe outputs from large language models.

DECENT: Distribution-aware Prototypical Calibration with Fine-grained Voting for VLM Misclassification Detection

Yijun Gong*, Bohan Yang*, Zhi Zhang, Yupeng Qin, Wenpeng Xing, Changting Lin, Meng Han, Ziyue Qiao#, Xiao Luo#

Under review Co-First Author

DECENT studies distribution-aware prototypical calibration and fine-grained voting for detecting misclassifications in vision-language models.

DECENT: Distribution-aware Prototypical Calibration with Fine-grained Voting for VLM Misclassification Detection

Yijun Gong*, Bohan Yang*, Zhi Zhang, Yupeng Qin, Wenpeng Xing, Changting Lin, Meng Han, Ziyue Qiao#, Xiao Luo#

Under review Co-First Author

DECENT studies distribution-aware prototypical calibration and fine-grained voting for detecting misclassifications in vision-language models.

UW-3DGS: Unveiling Underwater Scenes and Scattering Media through 3D Gaussian Splatting

Wenpeng Xing, Bohan Yang, Jie Chen#, Zaifeng Yang, Changting Lin, Meng Han#

Under review Student First Author

UW-3DGS investigates 3D Gaussian Splatting for reconstructing underwater scenes and scattering media.

UW-3DGS: Unveiling Underwater Scenes and Scattering Media through 3D Gaussian Splatting

Wenpeng Xing, Bohan Yang, Jie Chen#, Zaifeng Yang, Changting Lin, Meng Han#

Under review Student First Author

UW-3DGS investigates 3D Gaussian Splatting for reconstructing underwater scenes and scattering media.

All publications