With the continuous emergence of Chinese Large Language Models (LLMs), how to evaluate a model's capabilities has become an increasingly significant issue. The absence of a comprehensive Chinese benchmark that thoroughly assesses a model's performance, the unstandardized and incomparable prompting procedure, and the prevalent risk of contamination pose major challenges in the current evaluation of Chinese LLMs. We present CLEVA, a user-friendly platform crafted to holistically evaluate Chinese LLMs. Our platform employs a standardized workflow to assess LLMs' performance across various dimensions, regularly updating a competitive leaderboard. To alleviate contamination, CLEVA curates a significant proportion of new data and develops a sampling strategy that guarantees a unique subset for each leaderboard round. Empowered by an easy-to-use interface that requires just a few mouse clicks and a model API, users can conduct a thorough evaluation with minimal coding. Large-scale experiments featuring 23 Chinese LLMs have validated CLEVA's efficacy.

Abstract--Recently, zero-shot (or training-free) Neural Architecture Search (NAS) approaches have been proposed to liberate NAS from the expensive training process. The key idea behind zero-shot NAS approaches is to design proxies that can predict the accuracy of some given networks without training the network parameters. The proxies proposed so far are usually inspired by recent progress in theoretical understanding of deep learning and have shown great potential on several datasets and NAS benchmarks. This paper aims to comprehensively review and compare the state-of-the-art (SOTA) zero-shot NAS approaches, with an emphasis on their hardware awareness. To this end, we first review the mainstream zero-shot proxies and discuss their theoretical underpinnings. We then compare these zero-shot proxies through large-scale experiments and demonstrate their effectiveness in both hardware-aware and hardware-oblivious NAS scenarios. Finally, we point out several promising ideas to design better proxies. In recent years, deep neural networks have made significant via a hyper-network [11], [32], [33], [34], [35], [36], [37]. As breakthroughs in many applications, such as recommendation shown in Figure 2, one-shot NAS only needs to train a single systems, image classification, and natural language hyper-network instead of multiple candidate architectures modeling [1], [2], [3], [4], [5], [6], [7]. To automatically design whose number is usually exponentially large.

The ability to detect and analyze failed executions automatically is crucial for an explainable and robust robotic system. Recently, Large Language Models (LLMs) have demonstrated strong reasoning abilities on textual inputs. To leverage the power of LLMs for robot failure explanation, we introduce REFLECT, a framework which queries LLM for failure reasoning based on a hierarchical summary of robot past experiences generated from multisensory observations. The failure explanation can further guide a language-based planner to correct the failure and complete the task. To systematically evaluate the framework, we create the RoboFail dataset with a variety of tasks and failure scenarios. We demonstrate that the LLM-based framework is able to generate informative failure explanations that assist successful correction planning.

We present Generalized LoRA (GLoRA), an advanced approach for universal parameter-efficient fine-tuning tasks. Enhancing Low-Rank Adaptation (LoRA), GLoRA employs a generalized prompt module to optimize pre-trained model weights and adjust intermediate activations, providing more flexibility and capability across diverse tasks and datasets. Moreover, GLoRA facilitates efficient parameter adaptation by employing a scalable, modular, layer-wise structure search that learns individual adapter of each layer. Originating from a unified mathematical formulation, GLoRA exhibits strong transfer learning, few-shot learning and domain generalization abilities, as it adapts to new tasks through not only weights but also additional dimensions like activations. Comprehensive experiments demonstrate that GLoRA outperforms all previous methods in natural, specialized, and structured vision benchmarks, achieving superior accuracy with fewer parameters and computations. The proposed method on LLaMA-1 and LLaMA-2 also show considerable enhancements compared to the original LoRA in the language domain. Furthermore, our structural re-parameterization design ensures that GLoRA incurs no extra inference cost, rendering it a practical solution for resource-limited applications. Code and models are available at: GitHub. Large-scale deep neural networks have revolutionized the field of artificial intelligence, demonstrating unprecedented performance across various tasks and domains.

Foundation models are first pre-trained on vast unsupervised datasets and then fine-tuned on labeled data. Reinforcement learning, notably from human feedback (RLHF), can further align the network with the intended usage. Yet the imperfections in the proxy reward may hinder the training and lead to suboptimal results; the diversity of objectives in real-world tasks and human opinions exacerbate the issue. This paper proposes embracing the heterogeneity of diverse rewards by following a multi-policy strategy. Rather than focusing on a single a priori reward, we aim for Pareto-optimal generalization across the entire space of preferences. To this end, we propose rewarded soup, first specializing multiple networks independently (one for each proxy reward) and then interpolating their weights linearly. This succeeds empirically because we show that the weights remain linearly connected when fine-tuned on diverse rewards from a shared pre-trained initialization. We demonstrate the effectiveness of our approach for text-to-text (summarization, Q&A, helpful assistant, review), text-image (image captioning, text-to-image generation, visual grounding, VQA), and control (locomotion) tasks. We hope to enhance the alignment of deep models, and how they interact with the world in all its diversity.

Large Language Model (LLM) has gained popularity and achieved remarkable results in open-domain tasks, but its performance in real industrial domain-specific scenarios is average due to its lack of specific domain knowledge. This issue has attracted widespread attention, but there are few relevant benchmarks available. In this paper, we provide a benchmark Question Answering (QA) dataset named MSQA, centered around Microsoft products and IT technical problems encountered by customers. This dataset contains industry cloud-specific QA knowledge, an area not extensively covered in general LLMs, making it well-suited for evaluating methods aiming to enhance LLMs' domain-specific capabilities. In addition, we propose a new model interaction paradigm that can empower LLM to achieve better performance on domain-specific tasks where it is not proficient. Extensive experiments demonstrate that the approach following our method outperforms the commonly used LLM with retrieval methods. We make our source code and sample data available at: https://aka.ms/Microsoft_QA.

Language models learn a great quantity of factual information during pretraining, and recent work localizes this information to specific model weights like mid-layer MLP weights. In this paper, we find that we can change how a fact is stored in a model by editing weights that are in a different location than where existing methods suggest that the fact is stored. This is surprising because we would expect that localizing facts to specific model parameters would tell us where to manipulate knowledge in models, and this assumption has motivated past work on model editing methods. Specifically, we show that localization conclusions from representation denoising (also known as Causal Tracing) do not provide any insight into which model MLP layer would be best to edit in order to override an existing stored fact with a new one. This finding raises questions about how past work relies on Causal Tracing to select which model layers to edit. Next, we consider several variants of the editing problem, including erasing and amplifying facts. For one of our editing problems, editing performance does relate to localization results from representation denoising, but we find that which layer we edit is a far better predictor of performance. Our results suggest, counterintuitively, that better mechanistic understanding of how pretrained language models work may not always translate to insights about how to best change their behavior. Our code is available at https://github.com/google/belief-localization

Despite the recent progress in language generation models, their outputs may not always meet user expectations. In this work, we study whether informational feedback in natural language can be leveraged to improve generation quality and user preference alignment. To this end, we consider factual consistency in summarization, the quality that the summary should only contain information supported by the input documents, as the user-expected preference. We collect a high-quality dataset, DeFacto, containing human demonstrations and informational natural language feedback consisting of corrective instructions, edited summaries, and explanations with respect to the factual consistency of the summary. Using our dataset, we study three natural language generation tasks: (1) editing a summary by following the human feedback, (2) generating human feedback for editing the original summary, and (3) revising the initial summary to correct factual errors by generating both the human feedback and edited summary. We show that DeFacto can provide factually consistent human-edited summaries and further insights into summarization factual consistency thanks to its informational natural language feedback. We further demonstrate that fine-tuned language models can leverage our dataset to improve the summary factual consistency, while large language models lack the zero-shot learning ability in our proposed tasks that require controllable text generation.

We present Pre-trained Machine Reader (PMR), a novel method for retrofitting pre-trained masked language models (MLMs) to pre-trained machine reading comprehension (MRC) models without acquiring labeled data. PMR can resolve the discrepancy between model pre-training and downstream fine-tuning of existing MLMs. To build the proposed PMR, we constructed a large volume of general-purpose and high-quality MRC-style training data by using Wikipedia hyperlinks and designed a Wiki Anchor Extraction task to guide the MRC-style pre-training. Apart from its simplicity, PMR effectively solves extraction tasks, such as Extractive Question Answering and Named Entity Recognition. PMR shows tremendous improvements over existing approaches, especially in low-resource scenarios. When applied to the sequence classification task in the MRC formulation, PMR enables the extraction of high-quality rationales to explain the classification process, thereby providing greater prediction explainability. PMR also has the potential to serve as a unified model for tackling various extraction and classification tasks in the MRC formulation.

Enhancing the zero-shot performance of instruction-following models requires heavy computation, either by scaling the total number of training datasets or the model size. In this work, we explore how retrieval of soft prompts obtained through prompt tuning can efficiently assist hard prompts in zero-shot task generalization. Specifically, we train soft prompt embeddings for each prompt through prompt tuning, store the samples of the training instances mapped with the prompt embeddings, and retrieve the corresponding prompt embedding of the training instance closest to the query instance during inference. While only adding 0.007% additional parameters, retrieval of soft prompt enhances the performance of T0 on unseen tasks by outperforming it on 10 out of 11 datasets as well as improving the mean accuracy of T0 on BIG-bench benchmark by 2.39% points. Also, we report an interesting finding that retrieving source embeddings trained on similar answer choice formats is more important than those on similar task types.