Huanxuan Liao

@Xnhyacinth ยท User

GitHub profile โ†— ยท Compare

PhD student @ UCAS. Focusing on Long Context Modeling

Institute of Automation, Chinese Academy of SciencesBeijing93 followers28 repositories

Repositories

Xnhyacinth/OPERATE

Benchmarking Persistent Operational Agency in Source-Grounded Executable Systems

โ˜… 0PythonForks 0

Xnhyacinth/NesyCD

[AAAI 2025] Neural-Symbolic Collaborative Distillation: Advancing Small Language Models for Complex Reasoning Tasks

โ˜… 13PythonForks 2

Xnhyacinth/SKIntern

[COLING 2025] SKIntern: Internalizing Symbolic Knowledge for Distilling Better CoT Capabilities into Small Language Models

โ˜… 6PythonForks 0

Xnhyacinth/HyCo2

Beyond Hard and Soft: Hybrid Context Compression for Balancing Local and Global Information Retention

โ˜… 8PythonForks 0

Xnhyacinth/ResAdapt

ResAdapt: Adaptive Resolution for Efficient Multimodal Reasoning

โ˜… 10PythonForks 1

Xnhyacinth/SpaRTA

[ACL 2026 Main] Spectral Disentanglement: Rank-Aware Task Adaptation for Rehearsal-free Continual Learning in LLMs

โ˜… 9PythonForks 0

Xnhyacinth/threejs-games

Three.js games generated by multi-round AI benchmark: Arena Zero (FPS 5v5) + Rune Frontline (3D tower defense). Playable in browser.

โ˜… 1HTMLForks 0

Xnhyacinth/TAGI

[NeurIPS 2024] From Instance Training to Instruction Learning: Task Adapters Generation from Instructions

โ˜… 11PythonForks 0

Xnhyacinth/SparK

[AAAI 2026] SparK: Query-Aware Unstructured Sparsity with Recoverable KV Cache Channel Pruning

โ˜… 13PythonForks 1

Xnhyacinth/IAG

[COLING 2025] Awakening Augmented Generation: Learning to Awaken Internal Knowledge of Large Language Models for Question Answering

โ˜… 9PythonForks 2