Assessing foundation models’ abilities for human-level tasks is crucial for Artificial General Intelligence (AGI) development.Traditional benchmarks, which rely on artificial datasets, may not accurately represent these capabilities. In this paper, we introduce AGIEval, a novel bilingual benchmark designed to assess foundation models in the context of human-centric standardized exams, such as college entrance exams, law school admission tests, math competitions, and lawyer qualification tests. We evaluate several state-of-the-art foundation models on our benchmark. Impressively, we show that GPT-4 exceeds the average human performance in SAT, LSAT, and math contests, with 95% accuracy on SAT Math and 92.5% on the Chinese college entrance English exam. This demonstrates the exceptional performance of contemporary foundation models. In contrast, we also find that GPT-4 is less proficient in tasks requiring complex reasoning or specific domain knowledge. Our comprehensive analyses of model capabilities (understanding, knowledge, reasoning, and calculation) reveal their strengths and limitations, providing valuable insights into future directions for enhancing general capabilities. By concentrating on tasks pertinent to human cognition and decision-making, our benchmark delivers a meaningful and robust evaluation of foundation models’ performance in real-world scenarios.

Recent evaluations of Large Language Models (LLMs) have centered around testing their zero-shot/few-shot capabilities for basic natural language tasks and their ability to translate instructions into tool APIs. However, the evaluation of LLMs utilizing complex tools to finish multi-turn, multi-modal instructions in a complex multi-modal environment has not been investigated. To address this gap, we introduce the PowerPoint Task Completion (PPTC) benchmark to assess LLMs’ ability to create and edit PPT files based on user instructions. It contains 279 multi-turn sessions covering diverse topics and hundreds of instructions involving multi-modal operations. We also propose the PPTX-Match Evaluation System that evaluates if LLMs finish the instruction based on the prediction file rather than the label API sequence, thus it supports various LLM-generated API sequences. We measure 3 closed LLMs and 6 open-source LLMs. The results show that GPT-4 outperforms other LLMs with 75.1% accuracy in single-turn dialogue testing but faces challenges in completing entire sessions, achieving just 6% session accuracy. We find three main error causes in our benchmark: error accumulation in the multi-turn session, long PPT template processing, and multi-modality perception. These pose great challenges for future LLM and agent systems .

One major challenge for Large Language Models (LLMs) is completing complex tasks involving multiple entities, such as tool APIs. To tackle this, one approach is to retrieve relevant entities to enhance LLMs in task completion. A crucial issue here is obtaining accurate natural language representations for each entity to aid in retriever precision. In this paper, we propose the Natural Language Representation Optimization Problem, which aims to refine entity descriptions for improved retrieval and LLM utilization. We introduce the Learning to Represent with Natural Language method, which utilizes LLMs to optimize entity representations consisting of text patterns based on environmental feedback. We iteratively prompt LLMs to enhance or adjust patterns based on entity samples and evaluate their effectiveness through environmental feedback. Our method successfully learns human-understandable representations for classification tasks (e.g., instructions and documents) and API call tasks (e.g., APIbench and Virtual Home), significantly improving GPT-4’s task performance.

Large language models (LLMs) have demonstrated impressive reasoning capabilities, yet there is ongoing debate about these abilities and the potential data contamination problem recently. This paper aims to evaluate the reasoning capacities of LLMs, specifically in solving recent competition-level programming problems in Codeforces, which are expert-crafted and unique, requiring deep understanding and robust reasoning skills. We first provide a comprehensive evaluation of GPT-4’s perceived zero-shot performance on this task, considering various aspects such as problems’ release time, difficulties, and types of errors encountered. Surprisingly, the perceived performance of GPT-4 has experienced a cliff like decline in problems after September 2021 consistently across all the difficulties and types of problems, which shows the potential data contamination, as well as the challenges for any existing LLM to solve unseen complex reasoning problems. We further explore various approaches such as fine-tuning, Chain-of-Thought prompting and problem description simplification. Unfortunately, none of them is able to consistently mitigate the challenges. Through our work, we emphasize the importance of this excellent data source for assessing the genuine reasoning capabilities of LLMs, and foster the development of LLMs with stronger reasoning abilities and better generalization in the future.

Large Language Models (LLMs) have shown remarkable performance in various basic natural language tasks. For completing the complex task, we still need a plan for the task to guide LLMs to generate the specific solutions step by step. LLMs can directly generate task plans, but these plans may still contain factual errors or are incomplete. A high-quality task plan contains correct step-by-step solutions for solving all situations and behavioral instructions for avoiding mistakes. To obtain it, we propose the Learning to Plan method, which involves two phases: (1) In the first learning task plan phase, it iteratively updates the task plan with new step-by-step solutions and behavioral instructions, which are obtained by prompting LLMs to derive from training error feedback. (2) In the subsequent test phase, the LLM uses the learned task plan to guide the inference of LLM on the test set. We demonstrate the effectiveness of our method on the five different reasoning type tasks (8 datasets). Further, our analysis experiment shows that the task plan learned by one LLM can directly guide another LLM to improve its performance, which reveals a new transfer learning paradigm.

The growing dependence on Large Language Models (LLMs) for finishing user instructions necessitates a comprehensive understanding of their robustness to complex task completion in real-world situations. To address this critical need, we propose the PowerPoint Task Completion-Robustness (PPTC-R) benchmark to measure LLMs’ robustness to the user PPT task instruction and software version (Powerpoint). Specifically, we construct adversarial user instructions by attacking user instructions at sentence, semantic, and multi-language levels. To assess the robustness of Language Models to software versions, we vary the number of provided APIs to simulate both the newest version and earlier version settings. Subsequently, we test 3 closed-source and 4 open-source LLMs using a benchmark that incorporates these robustness settings, aiming to evaluate how deviations impact LLMs’ API calls for task completion. We find that GPT-4 exhibits the highest performance and strong robustness in our benchmark, particularly in the version update and the multilingual settings. However, we find that all LLMs lose their robustness when confronted with multiple challenges (e.g., multi-turn) simultaneously, leading to significant performance drops. We further analyze the robustness behavior and error reasons of LLMs in our benchmark, which provide valuable insights for researchers to understand the LLM’s robustness in task completion and develop more robust LLMs and agents.

Utilizing Large Language Models (LLMs) for complex tasks is challenging, often involving a time-consuming and uncontrollable prompt engineering process. This paper introduces a novel human-LLM interaction framework, Low-code LLM. It incorporates six types of simple low-code visual programming interactions to achieve more controllable and stable responses. Through visual interaction with a graphical user interface, users can incorporate their ideas into the process without writing trivial prompts. The proposed Low-code LLM framework consists of a Planning LLM that designs a structured planning workflow for complex tasks, which can be correspondingly edited and confirmed by users through low-code visual programming operations, and an Executing LLM that generates responses following the user-confirmed workflow. We highlight three advantages of the low-code LLM: user-friendly interaction, controllable generation, and wide applicability. We demonstrate its benefits using four typical applications. By introducing this framework, we aim to bridge the gap between humans and LLMs, enabling more effective and efficient utilization of LLMs for complex tasks. The code, prompts, and experimental details are available at https://github.com/moymix/TaskMatrix/tree/main/LowCodeLLM. A system demonstration video can be found at https://www.youtube.com/watch?v=jb2C1vaeO3E.

Existing research has shown that a multilingual pre-trained language model fine-tuned with one (source) language also performs well on downstream tasks for non-source languages, even though no fine-tuning is done on these languages. However, there is a clear gap between the performance of the source language and that of the non-source languages. This paper analyzes the fine-tuning process, discovers when the performance gap changes and identifies which network weights affect the overall performance most. Additionally, the paper seeks to answer to what extent the gap can be reduced by reducing forgetting. Based on the analysis results, a method named Fine-tuning slow and fast with four training policies is proposed to address these issues. Experimental results show the proposed method outperforms baselines by a clear margin.

Multilingual pre-trained language models, such as mBERT and XLM-R, have shown impressive cross-lingual ability. Surprisingly, both of them use multilingual masked language model (MLM) without any cross-lingual supervision or aligned data. Despite the encouraging results, we still lack a clear understanding of why cross-lingual ability could emerge from multilingual MLM. In our work, we argue that cross-language ability comes from the commonality between languages. Specifically, we study three language properties: constituent order, composition and word co-occurrence. First, we create an artificial language by modifying property in source language. Then we study the contribution of modified property through the change of cross-language transfer results on target language. We conduct experiments on six languages and two cross-lingual NLP tasks (textual entailment, sentence retrieval). Our main conclusion is that the contribution of constituent order and word co-occurrence is limited, while the composition is more crucial to the success of cross-linguistic transfer.

Dense retrieval has achieved impressive advances in first-stage retrieval from a large-scale document collection, which is built on bi-encoder architecture to produce single vector representation of query and document. However, a document can usually answer multiple potential queries from different views. So the single vector representation of a document is hard to match with multi-view queries, and faces a semantic mismatch problem. This paper proposes a multi-view document representation learning framework, aiming to produce multi-view embeddings to represent documents and enforce them to align with different queries. First, we propose a simple yet effective method of generating multiple embeddings through viewers. Second, to prevent multi-view embeddings from collapsing to the same one, we further propose a global-local loss with annealed temperature to encourage the multiple viewers to better align with different potential queries. Experiments show our method outperforms recent works and achieves state-of-the-art results.

Multilingual pre-trained models have demonstrated their effectiveness in many multilingual NLP tasks and enabled zero-shot or few-shot transfer from high-resource languages to low-resource ones. However, due to significant typological differences and contradictions between some languages, such models usually perform poorly on many languages and cross-lingual settings, which shows the difficulty of learning a single model to handle massive diverse languages well at the same time. To alleviate this issue, we present a new multilingual pre-training pipeline. We propose to generate language representation from multilingual pre-trained model and conduct linguistic analysis to show that language representation similarity reflects linguistic similarity from multiple perspectives, including language family, geographical sprachbund, lexicostatistics, and syntax. Then we cluster all the target languages into multiple groups and name each group as a representation sprachbund. Thus, languages in the same representation sprachbund are supposed to boost each other in both pre-training and fine-tuning as they share rich linguistic similarity. We pre-train one multilingual model for each representation sprachbund. Experiments are conducted on cross-lingual benchmarks and significant improvements are achieved compared to strong baselines.

Multilingual pre-trained models could leverage the training data from a rich source language (such as English) to improve performance on low resource languages. However, the transfer quality for multilingual Machine Reading Comprehension (MRC) is significantly worse than sentence classification tasks mainly due to the requirement of MRC to detect the word level answer boundary. In this paper, we propose two auxiliary tasks in the fine-tuning stage to create additional phrase boundary supervision: (1) A mixed MRC task, which translates the question or passage to other languages and builds cross-lingual question-passage pairs; (2) A language-agnostic knowledge masking task by leveraging knowledge phrases mined from web. Besides, extensive experiments on two cross-lingual MRC datasets show the effectiveness of our proposed approach.

Natural Questions is a new challenging machine reading comprehension benchmark with two-grained answers, which are a long answer (typically a paragraph) and a short answer (one or more entities inside the long answer). Despite the effectiveness of existing methods on this benchmark, they treat these two sub-tasks individually during training while ignoring their dependencies. To address this issue, we present a novel multi-grained machine reading comprehension framework that focuses on modeling documents at their hierarchical nature, which are different levels of granularity: documents, paragraphs, sentences, and tokens. We utilize graph attention networks to obtain different levels of representations so that they can be learned simultaneously. The long and short answers can be extracted from paragraph-level representation and token-level representation, respectively. In this way, we can model the dependencies between the two-grained answers to provide evidence for each other. We jointly train the two sub-tasks, and our experiments show that our approach significantly outperforms previous systems at both long and short answer criteria.

In this paper, we introduce XGLUE, a new benchmark dataset to train large-scale cross-lingual pre-trained models using multilingual and bilingual corpora, and evaluate their performance across a diverse set of cross-lingual tasks. Comparing to GLUE (Wang et al.,2019), which is labeled in English and includes natural language understanding tasks only, XGLUE has three main advantages: (1) it provides two corpora with different sizes for cross-lingual pre-training; (2) it provides 11 diversified tasks that cover both natural language understanding and generation scenarios; (3) for each task, it provides labeled data in multiple languages. We extend a recent cross-lingual pre-trained model Unicoder (Huang et al., 2019) to cover both understanding and generation tasks, which is evaluated on XGLUE as a strong baseline. We also evaluate the base versions (12-layer) of Multilingual BERT, XLM and XLM-R for comparison.

Understanding narrated instructional videos is important for both research and real-world web applications. Motivated by video dense captioning, we propose a model to generate procedure captions from narrated instructional videos which are a sequence of step-wise clips with description. Previous works on video dense captioning learn video segments and generate captions without considering transcripts. We argue that transcripts in narrated instructional videos can enhance video representation by providing fine-grained complimentary and semantic textual information. In this paper, we introduce a framework to (1) extract procedures by a cross-modality module, which fuses video content with the entire transcript; and (2) generate captions by encoding video frames as well as a snippet of transcripts within each extracted procedure. Experiments show that our model can achieve state-of-the-art performance in procedure extraction and captioning, and the ablation studies demonstrate that both the video frames and the transcripts are important for the task.

We present Unicoder, a universal language encoder that is insensitive to different languages. Given an arbitrary NLP task, a model can be trained with Unicoder using training data in one language and directly applied to inputs of the same task in other languages. Comparing to similar efforts such as Multilingual BERT and XLM , three new cross-lingual pre-training tasks are proposed, including cross-lingual word recovery, cross-lingual paraphrase classification and cross-lingual masked language model. These tasks help Unicoder learn the mappings among different languages from more perspectives. We also find that doing fine-tuning on multiple languages together can bring further improvement. Experiments are performed on two tasks: cross-lingual natural language inference (XNLI) and cross-lingual question answering (XQA), where XLM is our baseline. On XNLI, 1.8% averaged accuracy improvement (on 15 languages) is obtained. On XQA, which is a new cross-lingual dataset built by us, 5.5% averaged accuracy improvement (on French and German) is obtained.