Foundational challenges in assuring alignment and safety of large language models
This work identifies 18 foundational challenges in assuring the alignment and safety of large
language models (LLMs). These challenges are organized into three different categories …
language models (LLMs). These challenges are organized into three different categories …
Risk taxonomy, mitigation, and assessment benchmarks of large language model systems
Large language models (LLMs) have strong capabilities in solving diverse natural language
processing tasks. However, the safety and security issues of LLM systems have become the …
processing tasks. However, the safety and security issues of LLM systems have become the …
Generative AI security: challenges and countermeasures
arXiv:2402.12617v1 [cs.CR] 20 Feb 2024 Page 1 Generative AI Security: Challenges and
Countermeasures Banghua Zhu1, Norman Mu1, Jiantao Jiao1, and David Wagner1 1University …
Countermeasures Banghua Zhu1, Norman Mu1, Jiantao Jiao1, and David Wagner1 1University …
Watermark Smoothing Attacks against Language Models
Watermarking is a technique used to embed a hidden signal in the probability distribution of
text generated by large language models (LLMs), enabling attribution of the text to the …
text generated by large language models (LLMs), enabling attribution of the text to the …
Robust detection of watermarks for large language models under human edits
Watermarking has offered an effective approach to distinguishing text generated by large
language models (LLMs) from human-written text. However, the pervasive presence of …
language models (LLMs) from human-written text. However, the pervasive presence of …
SoK: Watermarking for AI-Generated Content
As the outputs of generative AI (GenAI) techniques improve in quality, it becomes
increasingly challenging to distinguish them from human-created content. Watermarking …
increasingly challenging to distinguish them from human-created content. Watermarking …
Data-Centric AI in the Age of Large Language Models
This position paper proposes a data-centric viewpoint of AI research, focusing on large
language models (LLMs). We start by making the key observation that data is instrumental in …
language models (LLMs). We start by making the key observation that data is instrumental in …
Position Paper: Data-Centric AI in the Age of Large Language Models
This position paper proposes a data-centric viewpoint of AI research, focusing on large
language models (LLMs). We start by making a key observation that data is instrumental in …
language models (LLMs). We start by making a key observation that data is instrumental in …
A Statistical Framework of Watermarks for Large Language Models: Pivot, Detection Efficiency and Optimal Rules
Since ChatGPT was introduced in November 2022, embedding (nearly) unnoticeable
statistical signals into text generated by large language models (LLMs), also known as …
statistical signals into text generated by large language models (LLMs), also known as …
A Watermark for Black-Box Language Models
Watermarking has recently emerged as an effective strategy for detecting the outputs of
large language models (LLMs). Most existing schemes require\emph {white-box} access to …
large language models (LLMs). Most existing schemes require\emph {white-box} access to …