Just months after OpenAI's ChatGPT chatbot upended the startup economy, cybercriminals and hackers are claiming to have created their own versions of the text-generating technology. The systems could, theoretically at least, supercharge criminals' ability to write malware or phishing emails that trick people into handing over their login information. Since the start of July, criminals posting on dark-web forums and marketplaces have been touting two large language models (LLMs) they say they've produced. The systems, which are said to mimic the functionalities of ChatGPT and Google's Bard, generate text to answer the questions or prompts users enter. But unlike the LLMs made by legitimate companies, these chatbots are marketed for illegal activities.

When trying to gain better visibility into a machine learning model in order to understand and mitigate the associated risks, a potentially valuable source of evidence is: which training examples most contribute to a given behavior? Influence functions aim to answer a counterfactual: how would the model's parameters (and hence its outputs) change if a given sequence were added to the training set? While influence functions have produced insights for small models, they are difficult to scale to large language models (LLMs) due to the difficulty of computing an inverse-Hessian-vector product (IHVP). We use the Eigenvalue-corrected Kronecker-Factored Approximate Curvature (EK-FAC) approximation to scale influence functions up to LLMs with up to 52 billion parameters. In our experiments, EK-FAC achieves similar accuracy to traditional influence function estimators despite the IHVP computation being orders of magnitude faster. We investigate two algorithmic techniques to reduce the cost of computing gradients of candidate training sequences: TF-IDF filtering and query batching. We use influence functions to investigate the generalization patterns of LLMs, including the sparsity of the influence patterns, increasing abstraction with scale, math and programming abilities, cross-lingual generalization, and role-playing behavior. Despite many apparently sophisticated forms of generalization, we identify a surprising limitation: influences decay to near-zero when the order of key phrases is flipped. Overall, influence functions give us a powerful new tool for studying the generalization properties of LLMs.

Given a pre-trained language model, how can we efficiently compress it without retraining? Retraining-free structured pruning algorithms are crucial in pre-trained language model compression due to their significantly reduced pruning cost and capability to prune large language models. However, existing retraining-free algorithms encounter severe accuracy degradation, as they fail to preserve the useful knowledge of pre-trained models. In this paper, we propose K-pruning (Knowledge-preserving pruning), an accurate retraining-free structured pruning algorithm for pre-trained language models. K-pruning identifies and prunes attention heads and neurons deemed to be superfluous, based on the amount of their inherent knowledge. K-pruning applies an iterative process of pruning followed by knowledge reconstruction for each sub-layer to preserve the knowledge of the pre-trained models. Consequently, K-pruning shows up to 58.02%p higher F1 score than existing retraining-free pruning algorithms under a high compression rate of 80% on the SQuAD benchmark.

The application of Artificial Intelligence (AI) in healthcare has been revolutionary, especially with the recent advancements in transformer-based Large Language Models (LLMs). However, the task of understanding unstructured electronic medical records remains a challenge given the nature of the records (e.g., disorganization, inconsistency, and redundancy) and the inability of LLMs to derive reasoning paradigms that allow for comprehensive understanding of medical variables. In this work, we examine the power of coupling symbolic reasoning with language modeling toward improved understanding of unstructured clinical texts. We show that such a combination improves the extraction of several medical variables from unstructured records. In addition, we show that the state-of-the-art commercially-free LLMs enjoy retrieval capabilities comparable to those provided by their commercial counterparts. Finally, we elaborate on the need for LLM steering through the application of symbolic reasoning as the exclusive use of LLMs results in the lowest performance.

The research explores the steerability of Large Language Models (LLMs), particularly OpenAI's ChatGPT iterations. By employing a behavioral psychology framework called OCEAN (Openness, Conscientiousness, Extroversion, Agreeableness, Neuroticism), we quantitatively gauged the model's responsiveness to tailored prompts. When asked to generate text mimicking an extroverted personality, OCEAN scored the language alignment to that behavioral trait. In our analysis, while "openness" presented linguistic ambiguity, "conscientiousness" and "neuroticism" were distinctly evoked in the OCEAN framework, with "extroversion" and "agreeableness" showcasing a notable overlap yet distinct separation from other traits. Our findings underscore GPT's versatility and ability to discern and adapt to nuanced instructions. Furthermore, historical figure simulations highlighted the LLM's capacity to internalize and project instructible personas, precisely replicating their philosophies and dialogic styles. However, the rapid advancements in LLM capabilities and the opaque nature of some training techniques make metric proposals degrade rapidly. Our research emphasizes a quantitative role to describe steerability in LLMs, presenting both its promise and areas for further refinement in aligning its progress to human intentions.

Creative coding tasks are often exploratory in nature. When producing digital artwork, artists usually begin with a high-level semantic construct such as a "stained glass filter" and programmatically implement it by varying code parameters such as shape, color, lines, and opacity to produce visually appealing results. Based on interviews with artists, it can be effortful to translate semantic constructs to program syntax, and current programming tools don't lend well to rapid creative exploration. To address these challenges, we introduce Spellburst, a large language model (LLM) powered creative-coding environment. Spellburst provides (1) a node-based interface that allows artists to create generative art and explore variations through branching and merging operations, (2) expressive prompt-based interactions to engage in semantic programming, and (3) dynamic prompt-driven interfaces and direct code editing to seamlessly switch between semantic and syntactic exploration. Our evaluation with artists demonstrates Spellburst's potential to enhance creative coding practices and inform the design of computational creativity tools that bridge semantic and syntactic spaces.

With ChatGPT-like large language models (LLM) prevailing in the community, how to evaluate the ability of LLMs is an open question. Existing evaluation methods suffer from following shortcomings: (1) constrained evaluation abilities, (2) vulnerable benchmarks, (3) unobjective metrics. We suggest that task-based evaluation, where LLM agents complete tasks in a simulated environment, is a one-for-all solution to solve above problems. We present AgentSims, an easy-to-use infrastructure for researchers from all disciplines to test the specific capacities they are interested in. Researchers can build their evaluation tasks by adding agents and buildings on an interactive GUI or deploy and test new support mechanisms, i.e. memory, planning and tool-use systems, by a few lines of codes. Our demo is available at https://agentsims.com .

Large Language Model (LLM) based Generative AI systems have seen significant progress in recent years. Integrating a knowledge retrieval architecture allows for seamless integration of private data into publicly available Generative AI systems using pre-trained LLM without requiring additional model fine-tuning. Moreover, Retrieval-Centric Generation (RCG) approach, a promising future research direction that explicitly separates roles of LLMs and retrievers in context interpretation and knowledge memorization, potentially leads to more Figure 1: Retrieval-Centric Generation (RCG) approach efficient implementation. SimplyRetrieve is an presents an innovative concept that leverages the mutually open-source tool with the goal of providing beneficial interaction between LLMs and retrievers a localized, lightweight, and user-friendly interface for more efficient context interpretation and knowledge to these sophisticated advancements to memorization. Increased clarity in role-separation between the machine learning community. SimplyRetrieve context interpretation and knowledge memorization features a GUI and API based RCG platform, can potentially boost the performance of generative assisted by a Private Knowledge Base AI systems.

Sycophancy is an undesirable behavior where models tailor their responses to follow a human user's view even when that view is not objectively correct (e.g., adapting liberal views once a user reveals that they are liberal). In this paper, we study the prevalence of sycophancy in language models and propose a simple synthetic-data intervention to reduce this behavior. First, on a set of three sycophancy tasks (Perez et al., 2022) where models are asked for an opinion on statements with no correct answers (e.g., politics), we observe that both model scaling and instruction tuning significantly increase sycophancy for PaLM models up to 540B parameters. Second, we extend sycophancy evaluations to simple addition statements that are objectively incorrect, finding that despite knowing that these statements are wrong, language models will still agree with them if the user does as well. To reduce sycophancy, we present a straightforward synthetic-data intervention that takes public NLP tasks and encourages models to be robust to user opinions on these tasks. Adding these data in a lightweight finetuning step can significantly reduce sycophantic behavior on held-out prompts. Code for generating synthetic data for intervention can be found at https://github.com/google/sycophancy-intervention.

Data management has traditionally relied on synthetic data generators to generate structured benchmarks, like the TPC suite, where we can control important parameters like data size and its distribution precisely. These benchmarks were central to the success and adoption of database management systems. But more and more, data management problems are of a semantic nature. An important example is finding tables that can be unioned. While any two tables with the same cardinality can be unioned, table union search is the problem of finding tables whose union is semantically coherent. Semantic problems cannot be benchmarked using synthetic data. Our current methods for creating benchmarks involve the manual curation and labeling of real data. These methods are not robust or scalable and perhaps more importantly, it is not clear how robust the created benchmarks are. We propose to use generative AI models to create structured data benchmarks for table union search. We present a novel method for using generative models to create tables with specified properties. Using this method, we create a new benchmark containing pairs of tables that are both unionable and non-unionable but related. We thoroughly evaluate recent existing table union search methods over existing benchmarks and our new benchmark. We also present and evaluate a new table search methods based on recent large language models over all benchmarks. We show that the new benchmark is more challenging for all methods than hand-curated benchmarks, specifically, the top-performing method achieves a Mean Average Precision of around 60%, over 30% less than its performance on existing manually created benchmarks. We examine why this is the case and show that the new benchmark permits more detailed analysis of methods, including a study of both false positives and false negatives that were not possible with existing benchmarks.