The advancement of large language models (LLMs) has led to a greater challenge of having a rigorous and systematic evaluation of complex tasks performed, especially in enterprise applications. Therefore, LLMs need to be able to benchmark enterprise datasets for various tasks. This work presents a systematic exploration of benchmarking strategies tailored to LLM evaluation, focusing on the utilization of domain-specific datasets and consisting of a variety of NLP tasks. The proposed evaluation framework encompasses 25 publicly available datasets from diverse enterprise domains like financial services, legal, cyber security, and climate and sustainability. The diverse performance of 13 models across different enterprise tasks highlights the importance of selecting the right model based on the specific requirements of each task. Code and prompts are available on GitHub.
As large language models (LLMs) continue to evolve, efficient evaluation metrics are vital for assessing their ability to compress information and reduce redundancy. While traditional metrics like Matrix Entropy offer valuable insights, they are computationally intensive for large-scale models due to their \( O(n^3) \) time complexity with Singular Value Decomposition (SVD). To mitigate this issue, we introduce the Matrix Nuclear-Norm, which not only serves as a metric to quantify the data compression proficiency of LLM but also provides a convex approximation of matrix rank to capture both predictive discriminability and diversity. By employing the \( L_{1,2}\text{-norm} \) to further approximate the nuclear norm, we can effectively assess the model's information compression capabilities. This approach reduces the time complexity to \( O(n^2) \) and eliminates the need for SVD computation. Consequently, the Matrix Nuclear-Norm achieves speeds 8 to 24 times faster than Matrix Entropy for the CEREBRAS-GPT model as sizes increase from 111M to 6.7B. This performance gap becomes more pronounced with larger models, as validated in tests with other models like Pythia. Additionally, evaluations on benchmarks and model responses confirm that our proposed Matrix Nuclear-Norm is a reliable, scalable, and efficient tool for assessing LLMs' performance, striking a balance between accuracy and computational efficiency. The code is available at https://github.com/MLGroupJLU/MatrixNuclearNorm.
We present a novel dataset aimed at advancing danger analysis and assessment by addressing the challenge of quantifying danger in video content and identifying how human-like a Large Language Model (LLM) evaluator is for the same. This is achieved by compiling a collection of 100 YouTube videos featuring various events. Each video is annotated by human participants who provided danger ratings on a scale from 0 (no danger to humans) to 10 (life-threatening), with precise timestamps indicating moments of heightened danger. Additionally, we leverage LLMs to independently assess the danger levels in these videos using video summaries. We introduce Mean Squared Error (MSE) scores for multimodal meta-evaluation of the alignment between human and LLM danger assessments. Our dataset not only contributes a new resource for danger assessment in video content but also demonstrates the potential of LLMs in achieving human-like evaluations.
Recently, there has been a trend of evaluating the Large Language Model (LLM) quality in the flavor of LLM-as-a-Judge, namely leveraging another LLM to evaluate the current output quality. However, existing judges are proven to be biased, namely they would favor answers which present better superficial quality (such as verbosity, fluency) while ignoring the instruction following ability. In this work, we propose systematic research about the bias of LLM-as-a-Judge. Specifically, for closed-source judge models, we apply calibration to mitigate the significance of superficial quality, both on probability level and prompt level. For open-source judge models, we propose to mitigate the bias by contrastive training, with curated negative samples that deviate from instruction but present better superficial quality. We apply our methods on the bias evaluation benchmark, and experiment results show our methods mitigate the bias by a large margin while maintaining a satisfactory evaluation accuracy.
Speech language models have recently demonstrated great potential as universal speech processing systems. Such models have the ability to model the rich acoustic information existing in audio signals, beyond spoken content, such as emotion, background noise, etc. Despite this, evaluation benchmarks which evaluate awareness to a wide range of acoustic aspects, are lacking. To help bridge this gap, we introduce SALMon, a novel evaluation suite encompassing background noise, emotion, speaker identity and room impulse response. The proposed benchmarks both evaluate the consistency of the inspected element and how much it matches the spoken text. We follow a modelling based approach, measuring whether a model gives correct samples higher scores than incorrect ones. This approach makes the benchmark fast to compute even for large models. We evaluated several speech language models on SALMon, thus highlighting the strengths and weaknesses of each evaluated method. We make the code and data publicly available at https://pages.cs.huji.ac.il/adiyoss-lab/salmon/ .
As large language models (LLMs) continue to evolve, the need for robust and standardized evaluation benchmarks becomes paramount. Evaluating the performance of these models is a complex challenge that requires careful consideration of various linguistic tasks, model architectures, and benchmarking methodologies. In recent years, various frameworks have emerged as noteworthy contributions to the field, offering comprehensive evaluation tests and benchmarks for assessing the capabilities of LLMs across diverse domains. This paper provides an exploration and critical analysis of some of these evaluation methodologies, shedding light on their strengths, limitations, and impact on advancing the state-of-the-art in natural language processing.
Developing prompt-based methods with Large Language Models (LLMs) requires making numerous decisions, which give rise to a combinatorial search problem. For example, selecting the right pre-trained LLM, prompt, and hyperparameters to attain the best performance for a task typically necessitates evaluating an expoential number of candidates on large validation sets. This exhaustive evaluation can be time-consuming and costly, as both inference and evaluation of LLM-based approaches are resource-intensive. Worse, a lot of computation is wasted: Many hyper-parameter settings are non-competitive, and many samples from the validation set are highly correlated - providing little or no new information. So, if the goal is to identify the best method, it can be done far more efficiently if the validation samples and methods are selected adaptively. In this paper, we propose a novel method to address this challenge. We lean on low-rank matrix factorization to fill in missing evaluations and on multi-armed bandits to sequentially identify the next (method, validation sample)-pair to evaluate. We carefully assess the efficacy of our approach on several competitive benchmark problems and show that it can identify the top-performing method using only 5-15% of the typically needed resources -- resulting in a staggering 85-95% LLM cost savings.
The training process of large language models (LLMs) often involves varying degrees of test data contamination. Although current LLMs are achieving increasingly better performance on various benchmarks, their performance in practical applications does not always match their benchmark results. Leakage of benchmarks can prevent the accurate assessment of LLMs' true performance. However, constructing new benchmarks is costly, labor-intensive and still carries the risk of leakage. Therefore, in this paper, we ask the question, Can we reuse these leaked benchmarks for LLM evaluation? We propose Inference-Time Decontamination (ITD) to address this issue by detecting and rewriting leaked samples without altering their difficulties. ITD can mitigate performance inflation caused by memorizing leaked benchmarks. Our proof-of-concept experiments demonstrate that ITD reduces inflated accuracy by 22.9% on GSM8K and 19.0% on MMLU. On MMLU, using Inference-time Decontamination can lead to a decrease in the results of Phi3 and Mistral by 6.7% and 3.6% respectively. We hope that ITD can provide more truthful evaluation results for large language models.
Prediction-powered inference (PPI) is a method that improves statistical estimates based on limited human-labeled data. PPI achieves this by combining small amounts of human-labeled data with larger amounts of data labeled by a reasonably accurate -- but potentially biased -- automatic system, in a way that results in tighter confidence intervals for certain parameters of interest (e.g., the mean performance of a language model). In this paper, we propose a method called Stratified Prediction-Powered Inference (StratPPI), in which we show that the basic PPI estimates can be considerably improved by employing simple data stratification strategies. Without making any assumptions on the underlying automatic labeling system or data distribution, we derive an algorithm for computing provably valid confidence intervals for population parameters (such as averages) that is based on stratified sampling. In particular, we show both theoretically and empirically that, with appropriate choices of stratification and sample allocation, our approach can provide substantially tighter confidence intervals than unstratified approaches. Specifically, StratPPI is expected to improve in cases where the performance of the autorater varies across different conditional distributions of the target data.
The acceleration of Large Language Models (LLMs) research has opened up new possibilities for evaluating generated texts. They serve as scalable and economical evaluators, but the question of how reliable these evaluators are has emerged as a crucial research question. Prior research efforts in the meta-evaluation of LLMs as judges limit the prompting of an LLM to a single use to obtain a final evaluation decision. They then compute the agreement between LLMs' outputs and human labels. This lacks interpretability in understanding the evaluation capability of LLMs. In light of this challenge, we propose Decompose and Aggregate, which breaks down the evaluation process into different stages based on pedagogical practices. Our experiments illustrate that it not only provides a more interpretable window for how well LLMs evaluate, but also leads to improvements up to 39.6% for different LLMs on a variety of meta-evaluation benchmarks.
While astonishingly capable, large Language Models (LLM) can sometimes produce outputs that deviate from human expectations. Such deviations necessitate an alignment phase to prevent disseminating untruthful, toxic, or biased information. Traditional alignment methods based on reinforcement learning often struggle with the identified instability, whereas preference optimization methods are limited by their overfitting to pre-collected hard-label datasets. In this paper, we propose a novel LLM alignment framework named $i$REPO, which utilizes implicit Reward pairwise difference regression for Empirical Preference Optimization. Particularly, $i$REPO employs self-generated datasets labeled by empirical human (or AI annotator) preference to iteratively refine the aligned policy through a novel regression-based loss function. Furthermore, we introduce an innovative algorithm backed by theoretical guarantees for achieving optimal results under ideal assumptions and providing a practical performance-gap result without such assumptions. Experimental results with Phi-2 and Mistral-7B demonstrate that $i$REPO effectively achieves self-alignment using soft-label, self-generated responses and the logit of empirical AI annotators. Furthermore, our approach surpasses preference optimization baselines in evaluations using the Language Model Evaluation Harness and Multi-turn benchmarks.
Effective evaluation of language models remains an open challenge in NLP. Researchers and engineers face methodological issues such as the sensitivity of models to evaluation setup, difficulty of proper comparisons across methods, and the lack of reproducibility and transparency. In this paper we draw on three years of experience in evaluating large language models to provide guidance and lessons for researchers. First, we provide an overview of common challenges faced in language model evaluation. Second, we delineate best practices for addressing or lessening the impact of these challenges on research. Third, we present the Language Model Evaluation Harness (lm-eval): an open source library for independent, reproducible, and extensible evaluation of language models that seeks to address these issues. We describe the features of the library as well as case studies in which the library has been used to alleviate these methodological concerns.
The rapid advancement of large language models has given rise to a plethora of applications across a myriad of real-world tasks, mainly centered on aligning with human intent. However, the complexities inherent in human intent necessitate a dependence on labor-intensive and time-consuming human evaluation. To alleviate this constraint, we delve into the paradigm of employing open-source large language models as evaluators, aligning with the prevailing trend of utilizing GPT-4. Particularly, we present a step-by-step evaluation framework: \textbf{Fennec}, capable of \textbf{F}ine-grained \textbf{E}valuatio\textbf{N} and correctio\textbf{N} \textbf{E}xtended through bran\textbf{C}hing and bridging. Specifically, the branching operation dissects the evaluation task into various dimensions and granularities, thereby alleviating the challenges associated with evaluation. Concurrently, the bridging operation amalgamates diverse training datasets, augmenting the variety of evaluation tasks. In experimental trials, our 7B model consistently outperforms open-source larger-scale evaluation models across various widely adopted benchmarks in terms of both \textit{Agreement} and \textit{Consistency}, closely approaching the capabilities of GPT-4. We employ the fine-grained correction capabilities induced by the evaluation model to refine multiple model responses, and the results show that the refinement elevates the quality of responses, leading to an improvement of 1-2 points on the MT-Bench. Our code is available at Github\footnote{\url{https://github.com/dropreg/Fennec}}.
A model's capacity to generalize its knowledge to interpret unseen inputs with different characteristics is crucial to build robust and reliable machine learning systems. Language model evaluation tasks lack information metrics about model generalization and their applicability in a new setting is measured using task and language-specific downstream performance, which is often lacking in many languages and tasks. In this paper, we explore a set of efficient and reliable measures that could aid in computing more information related to the generalization capability of language models in cross-lingual zero-shot settings. In addition to traditional measures such as variance in parameters after training and distance from initialization, we also measure the effectiveness of sharpness in loss landscape in capturing the success in cross-lingual transfer and propose a novel and stable algorithm to reliably compute the sharpness of a model optimum that correlates to generalization.
This paper studies the relationship between the surface form of a mathematical problem and its solvability by large language models. We find that subtle alterations in the surface form can significantly impact the answer distribution and the solve rate, exposing the language model's lack of robustness and sensitivity to the surface form in reasoning through complex problems. To improve mathematical reasoning performance, we propose Self-Consistency-over-Paraphrases (SCoP), which diversifies reasoning paths from specific surface forms of the problem. We evaluate our approach on four mathematics reasoning benchmarks over three large language models and show that SCoP improves mathematical reasoning performance over vanilla self-consistency, particularly for problems initially deemed unsolvable. Finally, we provide additional experiments and discussion regarding problem difficulty and surface forms, including cross-model difficulty agreement and paraphrasing transferability, and Variance of Variations (VOV) for language model evaluation.
This paper introduces Evalverse, a novel library that streamlines the evaluation of Large Language Models (LLMs) by unifying disparate evaluation tools into a single, user-friendly framework. Evalverse enables individuals with limited knowledge of artificial intelligence to easily request LLM evaluations and receive detailed reports, facilitated by an integration with communication platforms like Slack. Thus, Evalverse serves as a powerful tool for the comprehensive assessment of LLMs, offering both researchers and practitioners a centralized and easily accessible evaluation framework. Finally, we also provide a demo video for Evalverse, showcasing its capabilities and implementation in a two-minute format.
Large language model evaluation plays a pivotal role in the enhancement of its capacity. Previously, numerous methods for evaluating large language models have been proposed in this area. Despite their effectiveness, these existing works mainly focus on assessing objective questions, overlooking the capability to evaluate subjective questions which is extremely common for large language models. Additionally, these methods predominantly utilize centralized datasets for evaluation, with question banks concentrated within the evaluation platforms themselves. Moreover, the evaluation processes employed by these platforms often overlook personalized factors, neglecting to consider the individual characteristics of both the evaluators and the models being evaluated. To address these limitations, we propose a novel anonymous crowd-sourcing evaluation platform, BingJian, for large language models that employs a competitive scoring mechanism where users participate in ranking models based on their performance. This platform stands out not only for its support of centralized evaluations to assess the general capabilities of models but also for offering an open evaluation gateway. Through this gateway, users have the opportunity to submit their questions, testing the models on a personalized and potentially broader range of capabilities. Furthermore, our platform introduces personalized evaluation scenarios, leveraging various forms of human-computer interaction to assess large language models in a manner that accounts for individual user preferences and contexts. The demonstration of BingJian can be accessed at https://github.com/Mingyue-Cheng/Bingjian.
Despite their sophisticated capabilities, large language models (LLMs) encounter a major hurdle in effective assessment. This paper first revisits the prevalent evaluation method-multiple choice question answering (MCQA), which allows for straightforward accuracy measurement. Through a comprehensive evaluation of 24 models across 11 benchmarks, we highlight several potential drawbacks of MCQA, for instance, the inconsistency between the MCQA evaluation and the generation of open-ended responses in practical scenarios. In response, we introduce an RWQ-Elo rating system, engaging 24 LLMs such as GPT-4, GPT-3.5, Google-Gemini-Pro and LLaMA-1/-2, in a two-player competitive format, with GPT-4 serving as the judge. Each LLM receives an Elo rating thereafter. This system is designed to mirror real-world usage, and for this purpose, we have compiled a new benchmark called ``Real-world questions'' (RWQ), comprising 20,772 authentic user inquiries. Additionally, we thoroughly analyze the characteristics of our system and compare it with prior leaderboards like AlpacaEval and MT-Bench. Our analysis reveals the stability of our RWQ-Elo system, the feasibility of registering new models, and its potential to reshape LLM leaderboards.
Objective: This study aims to review the recent advances in community challenges for biomedical text mining in China. Methods: We collected information of evaluation tasks released in community challenges of biomedical text mining, including task description, dataset description, data source, task type and related links. A systematic summary and comparative analysis were conducted on various biomedical natural language processing tasks, such as named entity recognition, entity normalization, attribute extraction, relation extraction, event extraction, text classification, text similarity, knowledge graph construction, question answering, text generation, and large language model evaluation. Results: We identified 39 evaluation tasks from 6 community challenges that spanned from 2017 to 2023. Our analysis revealed the diverse range of evaluation task types and data sources in biomedical text mining. We explored the potential clinical applications of these community challenge tasks from a translational biomedical informatics perspective. We compared with their English counterparts, and discussed the contributions, limitations, lessons and guidelines of these community challenges, while highlighting future directions in the era of large language models. Conclusion: Community challenge evaluation competitions have played a crucial role in promoting technology innovation and fostering interdisciplinary collaboration in the field of biomedical text mining. These challenges provide valuable platforms for researchers to develop state-of-the-art solutions.
The focus of language model evaluation has transitioned towards reasoning and knowledge-intensive tasks, driven by advancements in pretraining large models. While state-of-the-art models are partially trained on large Arabic texts, evaluating their performance in Arabic remains challenging due to the limited availability of relevant datasets. To bridge this gap, we present \datasetname{}, the first multi-task language understanding benchmark for the Arabic language, sourced from school exams across diverse educational levels in different countries spanning North Africa, the Levant, and the Gulf regions. Our data comprises 40 tasks and 14,575 multiple-choice questions in Modern Standard Arabic (MSA) and is carefully constructed by collaborating with native speakers in the region. Our comprehensive evaluations of 35 models reveal substantial room for improvement, particularly among the best open-source models. Notably, BLOOMZ, mT0, LLaMA2, and Falcon struggle to achieve a score of 50%, while even the top-performing Arabic-centric model only achieves a score of 62.3%.