Enhancing Large Language Model Responses through Contrastive In-Context Learning
Leveraging contrastive examples, including positive and negative instances, can significantly improve the performance of large language models in generating responses that are better aligned with user preferences.
Plug and Play with Prompts: A Prompt Tuning Approach for Controlling Text Generation
A novel method, Plug and Play with Prompts (PPP), that utilizes prompt tuning to steer the generation of text by large language models in a data and parameter efficient manner.
Evaluating Decoding Strategies for Generating Coherent and Relevant Text with Pre-Trained GPT-2 Model
This research aims to comprehensively evaluate and compare various prominent decoding methodologies utilized in text generation with a pre-trained GPT-2 model. The study endeavors to establish a set of metrics to identify the most efficacious decoding technique, which can also serve as a tool for adversarial attacks on text classification models.
Iterative Refinement of Large Language Model Outputs Using Fine-Grained Actionable Feedback
LLMRefine, an inference-time optimization method, iteratively refines the output of large language models using a learned fine-grained feedback model to pinpoint defects and guide the refinement process.
Multi-Review Fusion-in-Context Study: Dataset, Evaluation, and Models
Decomposing grounded text generation tasks into subtasks, focusing on content fusion in a multi-document setting.
TEncDM: Understanding the Properties of Diffusion Model in Language Model Encodings
TEncDM introduces a novel approach for text generation using diffusion models trained in the latent space of language model encodings, showcasing superior performance over existing models.
Understanding Text Encoding Diffusion Model (TEncDM) for Text Generation
The author introduces TEncDM, a novel approach that utilizes language model encodings for text generation. By analyzing self-conditioning and decoder design, the author demonstrates the superiority of TEncDM over existing models.
© 2024 by Linnk AI