Open-sourced Large Language Models (LLMs) have achieved great success in various NLP tasks, however, they are still far inferior to API-based models when acting as agents. How to integrate agent ability into general LLMs becomes a crucial and urgent problem. This paper first delivers three key observations: (1) the current agent training corpus is entangled with both formats following and agent reasoning, which significantly shifts from the distribution of its pre-training data; (2) LLMs exhibit different learning speeds on the capabilities required by agent tasks; and (3) current approaches have side-effects when improving agent abilities by introducing hallucinations. Based on the above findings, we propose Agent-FLAN to effectively Fine-tune LANguage models for Agents. Through careful decomposition and redesign of the training corpus, Agent-FLAN enables Llama2-7B to outperform prior best works by 3.5\% across various agent evaluation datasets. With comprehensively constructed negative samples, Agent-FLAN greatly alleviates the hallucination issues based on our established evaluation benchmark. Besides, it consistently improves the agent capability of LLMs when scaling model sizes while slightly enhancing the general capability of LLMs. The code will be available at https://github.com/InternLM/Agent-FLAN.

In the era of flourishing large-scale models, the challenge of selecting and optimizing datasets from the vast and complex sea of data, to enhance the performance of large language models within the constraints of limited computational resources, has become paramount. This paper details our solution for the BetterMixture challenge, which focuses on the fine-tuning data mixing for large language models. Our approach, which secured third place, incorporates data deduplication, low-level and high-level quality filtering, and diversity selection. The foundation of our solution is Ke-Data-Juicer, an extension of Data-Juicer, demonstrating its robust capabilities in handling and optimizing data for large language models.

Recent advancements in Large Language Models (LLMs) have showcased remarkable capabilities across various tasks in different domains. However, the emergence of biases and the potential for generating harmful content in LLMs, particularly under malicious inputs, pose significant challenges. Current mitigation strategies, while effective, are not resilient under adversarial attacks. This paper introduces Resilient Guardrails for Large Language Models (RigorLLM), a novel framework designed to efficiently and effectively moderate harmful and unsafe inputs and outputs for LLMs. By employing a multi-faceted approach that includes energy-based training data augmentation through Langevin dynamics, optimizing a safe suffix for inputs via minimax optimization, and integrating a fusion-based model combining robust KNN with LLMs based on our data augmentation, RigorLLM offers a robust solution to harmful content moderation. Our experimental evaluations demonstrate that RigorLLM not only outperforms existing baselines like OpenAI API and Perspective API in detecting harmful content but also exhibits unparalleled resilience to jailbreaking attacks. The innovative use of constrained optimization and a fusion-based guardrail approach represents a significant step forward in developing more secure and reliable LLMs, setting a new standard for content moderation frameworks in the face of evolving digital threats.

Chinese Large Language Models (LLMs) have recently demonstrated impressive capabilities across various NLP benchmarks and real-world applications. However, the existing benchmarks for comprehensively evaluating these LLMs are still insufficient, particularly in terms of measuring knowledge that LLMs capture. Current datasets collect questions from Chinese examinations across different subjects and educational levels to address this issue. Yet, these benchmarks primarily focus on objective questions such as multiple-choice questions, leading to a lack of diversity in question types. To tackle this problem, we propose LHMKE, a Large-scale, Holistic, and Multi-subject Knowledge Evaluation benchmark in this paper. LHMKE is designed to provide a comprehensive evaluation of the knowledge acquisition capabilities of Chinese LLMs. It encompasses 10,465 questions across 75 tasks covering 30 subjects, ranging from primary school to professional certification exams. Notably, LHMKE includes both objective and subjective questions, offering a more holistic evaluation of the knowledge level of LLMs. We have assessed 11 Chinese LLMs under the zero-shot setting, which aligns with real examinations, and compared their performance across different subjects. We also conduct an in-depth analysis to check whether GPT-4 can automatically score subjective predictions. Our findings suggest that LHMKE is a challenging and advanced testbed for Chinese LLMs.

The paper underscores the significance of Large Language Models (LLMs) in reshaping recommender systems, attributing their value to unique reasoning abilities absent in traditional recommenders. Unlike conventional systems lacking direct user interaction data, LLMs exhibit exceptional proficiency in recommending items, showcasing their adeptness in comprehending intricacies of language. This marks a fundamental paradigm shift in the realm of recommendations. Amidst the dynamic research landscape, researchers actively harness the language comprehension and generation capabilities of LLMs to redefine the foundations of recommendation tasks. The investigation thoroughly explores the inherent strengths of LLMs within recommendation frameworks, encompassing nuanced contextual comprehension, seamless transitions across diverse domains, adoption of unified approaches, holistic learning strategies leveraging shared data reservoirs, transparent decision-making, and iterative improvements. Despite their transformative potential, challenges persist, including sensitivity to input prompts, occasional misinterpretations, and unforeseen recommendations, necessitating continuous refinement and evolution in LLM-driven recommender systems.

Watermarking approaches are proposed to identify if text being circulated is human or large language model (LLM) generated. The state-of-the-art watermarking strategy of Kirchenbauer et al. (2023a) biases the LLM to generate specific (``green'') tokens. However, determining the robustness of this watermarking method is an open problem. Existing attack methods fail to evade detection for longer text segments. We overcome this limitation, and propose {\em Self Color Testing-based Substitution (SCTS)}, the first ``color-aware'' attack. SCTS obtains color information by strategically prompting the watermarked LLM and comparing output tokens frequencies. It uses this information to determine token colors, and substitutes green tokens with non-green ones. In our experiments, SCTS successfully evades watermark detection using fewer number of edits than related work. Additionally, we show both theoretically and empirically that SCTS can remove the watermark for arbitrarily long watermarked text.

The task of financial analysis primarily encompasses two key areas: stock trend prediction and the corresponding financial question answering. Currently, machine learning and deep learning algorithms (ML&DL) have been widely applied for stock trend predictions, leading to significant progress. However, these methods fail to provide reasons for predictions, lacking interpretability and reasoning processes. Also, they can not integrate textual information such as financial news or reports. Meanwhile, large language models (LLMs) have remarkable textual understanding and generation ability. But due to the scarcity of financial training datasets and limited integration with real-time knowledge, LLMs still suffer from hallucinations and are unable to keep up with the latest information. To tackle these challenges, we first release AlphaFin datasets, combining traditional research datasets, real-time financial data, and handwritten chain-of-thought (CoT) data. It has a positive impact on training LLMs for completing financial analysis. We then use AlphaFin datasets to benchmark a state-of-the-art method, called Stock-Chain, for effectively tackling the financial analysis task, which integrates retrieval-augmented generation (RAG) techniques. Extensive experiments are conducted to demonstrate the effectiveness of our framework on financial analysis.

Large language models (LLMs) have significantly transformed the landscape of Natural Language Processing (NLP). Their impact extends across a diverse spectrum of tasks, revolutionizing how we approach language understanding and generations. Nevertheless, alongside their remarkable utility, LLMs introduce critical security and risk considerations. These challenges warrant careful examination to ensure responsible deployment and safeguard against potential vulnerabilities. This research paper thoroughly investigates security and privacy concerns related to LLMs from five thematic perspectives: security and privacy concerns, vulnerabilities against adversarial attacks, potential harms caused by misuses of LLMs, mitigation strategies to address these challenges while identifying limitations of current strategies. Lastly, the paper recommends promising avenues for future research to enhance the security and risk management of LLMs.

System-Level Test (SLT) has been a part of the test flow for integrated circuits for over a decade and still gains importance. However, no systematic approaches exist for test program generation, especially targeting non-functional properties of the Device under Test (DUT). Currently, test engineers manually compose test suites from off-the-shelf software, approximating the end-user environment of the DUT. This is a challenging and tedious task that does not guarantee sufficient control over non-functional properties. This paper proposes Large Language Models (LLMs) to generate test programs. We take a first glance at how pre-trained LLMs perform in test program generation to optimize non-functional properties of the DUT. Therefore, we write a prompt to generate C code snippets that maximize the instructions per cycle of a super-scalar, out-of-order architecture in simulation. Additionally, we apply prompt and hyperparameter optimization to achieve the best possible results without further training.

The rapid advancement of Generative Artificial Intelligence (GenAI) across diverse sectors raises significant environmental concerns, notably the carbon emissions from their cloud and high performance computing (HPC) infrastructure. This paper presents Sprout, an innovative framework designed to address these concerns by reducing the carbon footprint of generative Large Language Model (LLM) inference services. Sprout leverages the innovative concept of "generation directives" to guide the autoregressive generation process, thereby enhancing carbon efficiency. Our proposed method meticulously balances the need for ecological sustainability with the demand for high-quality generation outcomes. Employing a directive optimizer for the strategic assignment of generation directives to user prompts and an original offline quality evaluator, Sprout demonstrates a significant reduction in carbon emissions by over 40% in real-world evaluations using the Llama2 LLM and global electricity grid data. This research marks a critical step toward aligning AI technology with sustainable practices, highlighting the potential for mitigating environmental impacts in the rapidly expanding domain of generative artificial intelligence.