Papers by Sihao Liu

8 papers
H3Fusion: Helpful, Harmless, Honest Fusion of Aligned LLMs (2026.eacl-long)

Copied to clipboard

Challenge: Existing approaches to align pre-trained LLMs with instructions for one property are difficult to fine-tune.
Approach: They propose a mixture-of-experts-based fusion mechanism that models alignment as a controllable drift within the subspace, guided by a drift-regularization loss to balance competing alignment dimensions.
Outcome: Extensive evaluations of three benchmark datasets show that H3Fusion outperforms each individually aligned model by 11.37% and provides stronger robustness compared to the state-of-the-art LLM ensemble approaches by 13.77% and model-merging approaches by 6.18 %.
Design Challenges for a Multi-Perspective Search Engine (2022.findings-naacl)

Copied to clipboard

Challenge: a document retrieval system fails to deliver diverse and direct responses to controversial questions . classical document retrievals provide a ranked list of references to relevant but not necessarily trustworthy web documents .
Approach: They propose a perspective-oriented document retrieval paradigm to address these challenges . they propose sponses with different perspectives within topically-related web documents .
Outcome: The proposed system is based on a user survey and a prototype . it will be used to assess the utility and understanding of the system .
LLM-TOPLA: Efficient LLM Ensemble by Maximising Diversity (2024.findings-emnlp)

Copied to clipboard

Challenge: Extensive evaluation of modern large language models shows performance gain over component LLMs.
Approach: They propose a diversityoptimized LLM ensemble method with three unique properties . they introduce the focal diversity metric to capture diversityperformance correlation .
Outcome: The proposed method outperforms the best-performing ensemble on four benchmarks.
Talking Point based Ideological Discourse Analysis in News Events (2025.findings-acl)

Copied to clipboard

Challenge: Existing models of ideological discourse analysis fail to capture the key elements that shape real-world narratives and lack the ability to integrate contextual information required for understanding abstract ideological views.
Approach: They propose a framework motivated by the theory of ideological discourse analysis to analyze news articles related to real-world events.
Outcome: The proposed framework can generate ideology-specific viewpoints (partisan perspectives) it can be used to generate event snapshots, a visual way of interpreting event discourse.
Using LLM for Improving Key Event Discovery: Temporal-Guided News Stream Clustering with Event Summaries (2023.findings-emnlp)

Copied to clipboard

Challenge: Using hierarchical Dirichlet processes, we characterize news articles associated with key events from news streams.
Approach: They propose a generic framework for news stream clustering that analyzes the temporal trend of news articles to automatically extract the underlying key news events that draw significant media attention.
Outcome: The proposed framework produces more coherent clusters based on event summaries . the proposed framework is a first step in a new field of news analysis .
Evaluating Models’ Local Decision Boundaries via Contrast Sets (2020.findings-emnlp)

Copied to clipboard

Challenge: Standard test sets for supervised learning evaluate in-distribution generalization but are misleading when a dataset has systematic gaps.
Approach: They propose a more rigorous annotation paradigm for NLP that helps to close systematic gaps in the test data.
Outcome: The proposed model performs significantly lower on contrast sets than on the original test sets—up to 25% in some cases.
RetentiveKV: State-Space Memory for Uncertainty-Aware Multimodal KV Cache Eviction (2026.findings-acl)

Copied to clipboard

Challenge: Existing methods for evicting KV pairs rely on the "persistence of importance" hypothesis . visual tokens display "deferred importance" but become pivotal during later decoding, authors say .
Approach: They propose an entropy-driven method that reformulates KV eviction from "discrete context truncation" to "continuous memory evolution" they propose to prune visual tokens with "deferred importance" visual token exhibiting low salience but becoming pivotal during later decoding .
Outcome: The proposed method achieves 5.0 KV cache compression and 1.5 decoding acceleration.
MultiOpEd: A Corpus of Multi-Perspective News Editorials (2021.naacl-main)

Copied to clipboard

Challenge: a news editorial is a genre of persuasive text where argumentation structure is usually implicit.
Approach: They propose an open-domain news editorial corpus that supports automatic perspective discovery by identifying and abstracting natural language perspectives from editorials.
Outcome: The proposed system supports automatic perspective discovery tasks in news editorials.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations