Artificial intelligence (AI) systems have advanced significantly consequently of the introduction of Large Language Models (LLMs). Leading LLMs corresponding to ChatGPT released by OpenAI, Bard by Google, and Llama-2 have demonstrated their remarkable abilities in carrying out modern applications, starting from assisting in tool utilization and enhancing human evaluations to simulating human interactive behaviors. The extensive deployment of those LLMs has been made possible by their extraordinary competencies, however it comes with a big challenge of assuring the safety and dependability of their responses.

In relation to non-natural languages, specifically ciphers, recent research by a team has introduced several essential contributions that advance the understanding and application of LLMs. These innovations have been proposed with the aim of improving the dependability and safety of LLM interactions on this particular linguistic setting.

The team has introduced CipherChat, which is a framework created expressly to guage the applicability of safety alignment methods from the domain of natural languages to that of non-natural languages. In CipherChat, humans interact with LLMs through cipher-based prompts, detailed system role assignments, and succinct enciphered demonstrations. This architecture ensures that the LLMs’ understanding of ciphers, participation within the conversation, and sensitivity to inappropriate content are thoroughly examined.

This study highlights the critical need for the creation of safety alignment methods when working with non-natural languages, corresponding to ciphers, with a view to successfully match the capabilities of the underlying LLMs. While LLMs have shown extraordinary skill in understanding and producing human languages, the research says that additionally they exhibit unexpected prowess in comprehending non-natural languages. This information highlights the importance of developing safety regulations that cover these non-traditional types of communication in addition to people who fall inside the purview of traditional linguistics.

Quite a lot of experiments have been done using a wide range of realistic human ciphers on modern LLMs, corresponding to ChatGPT and GPT-4, to evaluate how well CipherChat performs. These evaluations cover 11 different safety topics and can be found in each Chinese and English. The findings point to a startling pattern which is that certain ciphers are capable of successfully get around GPT-4’s safety alignment procedures, with virtually 100% success rates in quite a lot of safety domains. This empirical result emphasizes the urgent necessity for creating customized safety alignment mechanisms for non-natural languages, like ciphers, to ensure the robustness and dependability of LLMs’ answers in various linguistic circumstances.

The team has shared that the research uncovers the phenomenon of the presence of a secret cipher inside LLMs. Drawing parallels to the concept of secret languages observed in other language models, the team has hypothesized that LLMs might possess a latent ability to decipher certain encoded inputs, thereby suggesting the existence of a singular cipher-related capability. 

Constructing on this statement, a singular and effective framework often known as SelfCipher has been introduced, which relies solely on role-play scenarios and a limited variety of demonstrations in natural language to tap into and activate the latent secret cipher capability inside LLMs. The efficacy of SelfCipher demonstrates the potential of harnessing these hidden abilities to reinforce LLM performance in deciphering encoded inputs and generating meaningful responses.

Take a look at the Paper, Project, and GitHub. All Credit For This Research Goes To the Researchers on This Project. Also, don’t forget to hitch our 28k+ ML SubReddit, 40k+ Facebook Community, Discord Channel, and Email Newsletter, where we share the most recent AI research news, cool AI projects, and more.