Language models can distinguish between testing and deployment phases -- a capability known as evaluation awareness. This has significant safety and policy implications, potentially undermining the reliability of evaluations that are central to AI governance frameworks and voluntary industry commitments. In this paper, we study evaluation awareness in Llama-3.3-70B-Instruct. We show that linear probes can separate real-world evaluation and deployment prompts, suggesting that current models internally represent this distinction. We also find that current safety evaluations are correctly classified by the probes, suggesting that they already appear artificial or inauthentic to models. Our findings underscore the importance of ensuring trustworthy evaluations and understanding deceptive capabilities. More broadly, our work showcases how model internals may be leveraged to support blackbox methods in safety audits, especially for future models more competent at evaluation awareness and deception.

Scientific research heavily depends on suitable datasets for method validation, but existing academic platforms with dataset management like PapersWithCode suffer from inefficiencies in their manual workflow. To overcome this bottleneck, we present a system, called ChatPD, that utilizes Large Language Models (LLMs) to automate dataset information extraction from academic papers and construct a structured paper-dataset network. Our system consists of three key modules: \textit{paper collection}, \textit{dataset information extraction}, and \textit{dataset entity resolution} to construct paper-dataset networks. Specifically, we propose a \textit{Graph Completion and Inference} strategy to map dataset descriptions to their corresponding entities. Through extensive experiments, we demonstrate that ChatPD not only outperforms the existing platform PapersWithCode in dataset usage extraction but also achieves about 90\% precision and recall in entity resolution tasks. Moreover, we have deployed ChatPD to continuously extract which datasets are used in papers, and provide a dataset discovery service, such as task-specific dataset queries and similar dataset recommendations. We open source ChatPD and the current paper-dataset network on this [GitHub repository]{https://github.com/ChatPD-web/ChatPD}.

Real-world enterprise text-to-SQL workflows often involve complex cloud or local data across various database systems, multiple SQL queries in various dialects, and diverse operations from data transformation to analytics. We introduce Spider 2.0, an evaluation framework comprising 632 real-world text-to-SQL workflow problems derived from enterprise-level database use cases. The databases in Spider 2.0 are sourced from real data applications, often containing over 1,000 columns and stored in local or cloud database systems such as BigQuery and Snowflake. We show that solving problems in Spider 2.0 frequently requires understanding and searching through database metadata, dialect documentation, and even project-level codebases. This challenge calls for models to interact with complex SQL workflow environments, process extremely long contexts, perform intricate reasoning, and generate multiple SQL queries with diverse operations, often exceeding 100 lines, which goes far beyond traditional text-to-SQL challenges. Our evaluations indicate that based on o1-preview, our code agent framework successfully solves only 17.0% of the tasks, compared with 91.2% on Spider 1.0 and 73.0% on BIRD. Our results on Spider 2.0 show that while language models have demonstrated remarkable performance in code generation -- especially in prior text-to-SQL benchmarks -- they require significant improvement in order to achieve adequate performance for real-world enterprise usage. Progress on Spider 2.0 represents crucial steps towards developing intelligent, autonomous, code agents for real-world enterprise settings. Our code, baseline models, and data are available at https://spider2-sql.github.io.

Out-of-distribution (OOD) detection aims to identify OOD inputs from unknown classes, which is important for the reliable deployment of machine learning models in the open world. Various scoring functions are proposed to distinguish it from in-distribution (ID) data. However, existing methods generally focus on excavating the discriminative information from a single input, which implicitly limits its representation dimension. In this work, we introduce a novel perspective, i.e., employing different common corruptions on the input space, to expand that. We reveal an interesting phenomenon termed confidence mutation, where the confidence of OOD data can decrease significantly under the corruptions, while the ID data shows a higher confidence expectation considering the resistance of semantic features. Based on that, we formalize a new scoring method, namely, Confidence aVerage (CoVer), which can capture the dynamic differences by simply averaging the scores obtained from different corrupted inputs and the original ones, making the OOD and ID distributions more separable in detection tasks. Extensive experiments and analyses have been conducted to understand and verify the effectiveness of CoVer.

Machine learning is widely utilized across various industries. Identifying the appropriate machine learning models and datasets for specific tasks is crucial for the effective industrial application of machine learning. However, this requires expertise in both machine learning and the relevant domain, leading to a high learning cost. Therefore, research focused on extracting combinations of tasks, machine learning models, and datasets from academic papers is critically important, as it can facilitate the automatic recommendation of suitable methods. Conventional information extraction methods from academic papers have been limited to identifying machine learning models and other entities as named entities. To address this issue, this study proposes a methodology extracting tasks, machine learning methods, and dataset names from scientific papers and analyzing the relationships between these information by using LLM, embedding model, and network clustering. The proposed method's expression extraction performance, when using Llama3, achieves an F-score exceeding 0.8 across various categories, confirming its practical utility. Benchmarking results on financial domain papers have demonstrated the effectiveness of this method, providing insights into the use of the latest datasets, including those related to ESG (Environmental, Social, and Governance) data.

As large language models (LLMs) continue to grow by scaling laws, reinforcement learning from human feedback (RLHF) has gained significant attention due to its outstanding performance. However, unlike pretraining or fine-tuning a single model, scaling reinforcement learning from human feedback (RLHF) for training large language models poses coordination challenges across four models. We present OpenRLHF, an open-source framework enabling efficient RLHF scaling. Unlike existing RLHF frameworks that co-locate four models on the same GPUs, OpenRLHF re-designs scheduling for the models beyond 70B parameters using Ray, vLLM, and DeepSpeed, leveraging improved resource utilization and diverse training approaches. Integrating seamlessly with Hugging Face, OpenRLHF provides an out-of-the-box solution with optimized algorithms and launch scripts, which ensures user-friendliness. OpenRLHF implements RLHF, DPO, rejection sampling, and other alignment techniques. Empowering state-of-the-art LLM development, OpenRLHF's code is available at https://github.com/OpenLLMAI/OpenRLHF.

We perform a missing, reproducible evaluation of all publicly available GPT-4 family models concerning the Document Understanding field, where it is frequently required to comprehend text spacial arrangement and visual clues in addition to textual semantics. Benchmark results indicate that though it is hard to achieve satisfactory results with text-only models, GPT-4 Vision Turbo performs well when one provides both text recognized by an external OCR engine and document images on the input. Evaluation is followed by analyses that suggest possible contamination of textual GPT-4 models and indicate the significant performance drop for lengthy documents.

Out-of-distribution (OOD) detection aims at identifying samples from unknown classes, playing a crucial role in trustworthy models against errors on unexpected inputs. Extensive research has been dedicated to exploring OOD detection in the vision modality. Vision-language models (VLMs) can leverage both textual and visual information for various multi-modal applications, whereas few OOD detection methods take into account information from the text modality. In this paper, we propose a novel post hoc OOD detection method, called NegLabel, which takes a vast number of negative labels from extensive corpus databases. We design a novel scheme for the OOD score collaborated with negative labels. Theoretical analysis helps to understand the mechanism of negative labels. Extensive experiments demonstrate that our method NegLabel achieves state-ofthe-art performance on various OOD detection benchmarks and generalizes well on multiple VLM architectures. Furthermore, our method NegLabel exhibits remarkable robustness against diverse domain shifts. In open-world scenarios, deploying machine learning models faces a critical challenge: how to handle data from unknown classes, commonly referred to as out-of-distribution (OOD) data (Hendrycks & Gimpel, 2017). The presence of OOD data can lead to models exhibiting overconfidence, potentially resulting in severe errors or security risks. This issue is particularly pronounced in critical applications, such as autonomous vehicles and medical diagnosis. Therefore, detecting and rejecting OOD data plays a crucial role in ensuring the reliability and safety of the model. Traditional visual OOD detection methods (Hsu et al., 2020a; Wang et al., 2021b; Huang et al., 2021; Sun et al., 2021; Wang et al., 2021a) typically rely solely on image information, ignoring the rich textual information carried by labels. Vision-language models (VLMs) can leverage multimodal information, which is also beneficial for OOD detection. Some recently proposed methods attempt to design dedicated OOD detectors for VLMs. Specifically, ZOC (Esmaeilpour et al., 2022) defines the new task - zero-shot OOD detection, and uses a trainable captioner to generate candidate OOD labels to match OOD images. However, when dealing with large-scale datasets encompassing a multitude of in-distribution (ID) classes, like ImageNet-1k, the captioner may not generate effective candidate OOD labels, resulting in poor performance. MCM (Ming et al., 2022a) uses the maximum logit of scaled softmax to identify OOD images. However, MCM only employs information from the ID label space and does not effectively exploit the text interpretation capabilities of VLMs.

Nonparametric learning is a fundamental concept in machine learning that aims to capture complex patterns and relationships in data without making strong assumptions about the underlying data distribution. Owing to simplicity and familiarity, one of the most well-known algorithms under this paradigm is the $k$-nearest neighbors ($k$-NN) algorithm. Driven by the usage of machine learning in safety-critical applications, in this work, we shed new light on the traditional nearest neighbors algorithm from the perspective of information theory and propose a robust and interpretable framework for tasks such as classification, regression, and anomaly detection using a single model. Instead of using a traditional distance measure which needs to be scaled and contextualized, we use a novel formulation of \textit{surprisal} (amount of information required to explain the difference between the observed and expected result). Finally, we demonstrate this architecture's capability to perform at-par or above the state-of-the-art on classification, regression, and anomaly detection tasks using a single model with enhanced interpretability by providing novel concepts for characterizing data and predictions.

The recognition of dataset names is a critical task for automatic information extraction in scientific literature, enabling researchers to understand and identify research opportunities. However, existing corpora for dataset mention detection are limited in size and naming diversity. In this paper, we introduce the Dataset Mentions Detection Dataset (DMDD), the largest publicly available corpus for this task. DMDD consists of the DMDD main corpus, comprising 31,219 scientific articles with over 449,000 dataset mentions weakly annotated in the format of in-text spans, and an evaluation set, which comprises of 450 scientific articles manually annotated for evaluation purposes. We use DMDD to establish baseline performance for dataset mention detection and linking. By analyzing the performance of various models on DMDD, we are able to identify open problems in dataset mention detection. We invite the community to use our dataset as a challenge to develop novel dataset mention detection models.