research-papers

Who Validates the Validators? Aligning LLM-Assisted Evaluation of LLM Outputs with Human Preferences

Original Paper: https://arxiv.org/abs/2404.12272 By: Shreya Shankar, J.D. Zamfirescu-Pereira, Björn Hartmann, Aditya G. Parameswaran, Ian Arawjo Abstract: Due to the cumbersome nature of human evaluation and limitations of code-based evaluation, Large Language Models (LLMs) are increasingly being used to assist humans in evaluating LLM outputs.

research-papers

CYBERSECEVAL 2: A Wide-Ranging Cybersecurity Evaluation Suite for Large Language Models

Original Paper: https://ai.meta.com/research/publications/cyberseceval-2-a-wide-ranging-cybersecurity-evaluation-suite-for-large-language-models/ By: Manish Bhatt∗, Sahana Chennabasappa∗, Yue Li∗, Cyrus Nikolaidis∗, Daniel Song∗, Shengye Wan∗, Faizan Ahmad, Cornelius Aschermann, Yaohui Chen, Dhaval Kapil, David Molnar, Spencer Whitman, Joshua Saxe∗ ∗Co-equal primary author Abstract: Large language models (LLMs) introduce new security risks, but there

research-papers

WizardLM: Empowering Large Language Models to Follow Complex Instructions

Original Paper: https://arxiv.org/abs/2304.12244 By: Can Xu, Qingfeng Sun, Kai Zheng, Xiubo Geng, Pu Zhao, Jiazhan Feng, Chongyang Tao, Daxin Jiang Abstract: Training large language models (LLMs) with open-domain instruction following data brings colossal success. However, manually creating such instruction data is very time-consuming and labor-intensive.

research-papers

On the Empirical Complexity of Reasoning and Planning in LLMs

Original Paper: https://arxiv.org/abs/2404.11041 By: Liwei Kang, Zirui Zhao, David Hsu, Wee Sun Lee Abstract: Large Language Models (LLMs) work surprisingly well for some complex reasoning problems via chain-of-thought (CoT) or tree-of-thought (ToT), but the underlying reasons remain unclear. We seek to understand the performance of

research-papers

How to Use a Custom Grading Criteria to Evaluate LLM Responses (LLM-as-a-Judge)

In the rapidly evolving field of language models, ensuring the accuracy and relevance of responses is crucial. This blog post will guide you through setting up a custom grading criteria to evaluate responses from large language models (LLMs), using a simple conditional evaluation system. What is it? A custom grading

research-papers

From Noise to Clarity: Unraveling the Adversarial Suffix of Large Language Model Attacks via Translation of Text Embeddings

Original Paper: https://arxiv.org/abs/2402.16006 By: Hao Wang, Hao Li, Minlie Huang, Lei Sha Abstract: The safety defense methods of Large language models(LLMs) stays limited because the dangerous prompts are manually curated to just few known attack types, which fails to keep pace with emerging varieties.

research-papers

Prompt Stealing Attacks Against Text-to-Image Generation Models

Original Paper: https://arxiv.org/abs/2302.09923 By: Xinyue Shen, Yiting Qu, Michael Backes, Yang Zhang Abstract: Text-to-Image generation models have revolutionized the artwork design process and enabled anyone to create high-quality images by entering text descriptions called prompts. Creating a high-quality prompt that consists of a subject and

research-papers

Ever: Mitigating Hallucination in Large Language Models through Real-Time Verification and Rectification

Original Paper: https://arxiv.org/html/2311.09114v2 By: Haoqiang Kang, Juntong Ni, Huaxiu Yao Abstract: Large Language Models (LLMs) have demonstrated remarkable proficiency in generating fluent text. However, they often encounter the challenge of generating inaccurate or hallucinated content. This issue is common in both non-retrieval-based generation and retrieval-augmented

research-papers

H2O-Danube-1.8B Technical Report

Original Paper: https://arxiv.org/abs/2401.16818 By: Philipp Singer, Pascal Pfeiffer, Yauhen Babakhin, Maximilian Jeblick, Nischay Dhankhar, Gabor Fodor, Sri Satish Ambati Abstract: We present H2O-Danube, a series of small 1.8B language models consisting of H2O-Danube-1.8B, trained on 1T tokens, and the incremental improved H2O-Danube2-1.8B

research-papers

Universal and Transferable Adversarial Attacks on Aligned Language Models

Original Paper: https://arxiv.org/abs/2307.15043 By: Andy Zou, Zifan Wang, Nicholas Carlini, Milad Nasr, J. Zico Kolter, Matt Fredrikson Abstract: Because "out-of-the-box" large language models are capable of generating a great deal of objectionable content, recent work has focused on aligning these models in an

research-papers

Post-Semantic-Thinking: A Robust Strategy to Distill Reasoning Capacity from Large Language Models

Original Paper: https://arxiv.org/html/2404.09170v1 By: Xiaoshu Chen, Sihang Zhou, Liang Ke, XinwangLiu Abstract Chain of thought finetuning aims to endow small student models with reasoning capacity to improve their performance towards a specific task by allowing them to imitate the reasoning procedure of large language models

research-papers

Direct Preference Optimization: Your Language Model is Secretly a Reward Model

Original Paper: https://arxiv.org/html/2305.18290v2 By: Rafael Rafailov, Archit Sharma, Eric Mitchell, Stefano Ermon, Christopher D. Manning, Chelsea Finn Abstract: While large-scale unsupervised language models (LMs) learn broad world knowledge and some reasoning skills, achieving precise control of their behavior is difficult due to the completely unsupervised