Mechanisms of Non-Factual Hallucination in Language Models: Investigating Causes and Detection
Understanding the causes of non-factual hallucinations in language models and proposing effective detection methods.
Investigating Bias in Large Language Models (LLMs) for Media Bias Detection
LLMs exhibit biases impacting media bias detection, requiring debiasing strategies for more equitable AI systems.
Large Language Models: Understanding Knowledge Cutoffs
LLMs' effective cutoff dates may differ from reported dates due to deduplication issues and outdated data sources.
OpenEval: Comprehensive Evaluation of Chinese LLMs
OpenEval introduces a comprehensive evaluation platform for Chinese LLMs, focusing on capability, alignment, and safety.
Safety Concerns in Large Language Models: Lessons from LLaMAs
Training large language models to follow instructions can lead to safety vulnerabilities, but adding safety examples during fine-tuning can significantly improve model safety without compromising performance.
Understanding Knowledge Cutoffs in Large Language Models
Effective knowledge cutoffs in LLMs differ from reported cutoff dates due to deduplication issues and temporal misalignment of CommonCrawl dumps.
Amortizing Intractable Inference in Large Language Models: A Bayesian Approach
Amortized Bayesian inference using GFlowNets enables efficient sampling from intractable posteriors in large language models.
Speculative Contrastive Decoding: Improving Language Model Inference Efficiency and Quality
Introducing Speculative Contrastive Decoding (SCD) to enhance decoding efficiency and quality in large language models by leveraging smaller models.
CoIN: A Benchmark of Continual Instruction Tuning for Multimodal Large Language Model
The author introduces CoIN as a benchmark to evaluate Multimodal Large Language Models in sequential instruction tuning, highlighting the importance of aligning with human intent and retaining knowledge.
Efficient Constrained Generation for Large Language Models
The author presents DOMINO, a novel decoding algorithm that achieves efficient and minimally invasive constrained generation, outperforming existing approaches with no loss in accuracy.
© 2024 by Linnk AI