TripPattern uses a three-way vocabulary split and an alternation pattern between two groups to watermark LLM output, preserving fluency better than two-group schemes.
Element-aware Summarization with Large Language Models: Expert-aligned Evaluation and Chain-of-Thought Method
5 Pith papers cite this work, alongside 41 external citations. Polarity classification is still indexing.
5
Pith papers citing it
41
external citations · external index
representative citing papers
Introduces the MEA benchmark for multi-target cross-lingual summarization across 24 languages and demonstrates that activation steering from English summarization representations improves performance.
A visual analytics workspace uses LLM explanation and verification agents to semi-automatically annotate and perturbation-check mapper graph elements of BERT embeddings, replicating known layer-wise linguistic patterns.
SHIELD shows that standard AUROC overstates AI-text detector quality, and that six zero-shot detectors collapse under a controllable word-replacement humanification.
A reasoning-distillation plus dual-reward GRPO method for multi-role dialogue summarization matches ROUGE and BERTScore baselines while improving factual faithfulness and preference alignment on CSDS and SAMSum.
citing papers explorer
-
TripPattern: A Pattern-based Text Watermarking Method for Large Language Models
TripPattern uses a three-way vocabulary split and an alternation pattern between two groups to watermark LLM output, preserving fluency better than two-group schemes.
-
Understanding LLM Behavior in Multi-Target Cross-Lingual Summarization
Introduces the MEA benchmark for multi-target cross-lingual summarization across 24 languages and demonstrates that activation steering from English summarization representations improves performance.
-
Explainable Mapper: Charting LLM Embedding Spaces Using Perturbation-Based Explanation and Verification Agents
A visual analytics workspace uses LLM explanation and verification agents to semi-automatically annotate and perturbation-check mapper graph elements of BERT embeddings, replicating known layer-wise linguistic patterns.
-
Beyond Easy Wins: A Text Hardness-Aware Benchmark for LLM-generated Text Detection
SHIELD shows that standard AUROC overstates AI-text detector quality, and that six zero-shot detectors collapse under a controllable word-replacement humanification.
-
Beyond Overlap Metrics: Rewarding Reasoning and Preferences for Faithful Multi-Role Dialogue Summarization
A reasoning-distillation plus dual-reward GRPO method for multi-role dialogue summarization matches ROUGE and BERTScore baselines while improving factual faithfulness and preference alignment on CSDS and SAMSum.