Thailand

Evaluation and ranking of large language models (LLMs) has become an important problem with the proliferation of these models and their impact. Evaluation methods either require human responses which are expensive to acquire or use pairs of LLMs to evaluate each other which can be unreliable. In this paper, we provide a novel perspective where, given a dataset of prompts (viz. questions, instructions, etc.) and a set of LLMs, we rank them without access to any ground truth or reference responses. Inspired by real life where both an expert and a knowledgeable person can identify a novice our main idea is to consider triplets of models, where each one of them evaluates the other two, correctly identifying the worst model in the triplet with high probability. We also analyze our idea and provide sufficient conditions for it to succeed. Applying this idea repeatedly we propose two methods to rank LLMs. In experiments on different generative tasks (summarization, multiple-choice, and dialog), our methods reliably recover true rankings without reference data. This points to a viable low-resource mechanism for practical use.

ACL 2024

Ranking Large Language Models without Ground Truth

llm-as-a-judge

ranking

large language models

benchmarking

poster

### Welcome!
The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. Our Virtual Poster Sessions will take place online Thursday, August 22, 2024.

You are required to register for this event. **Please register [here](https://2024.aclweb.org/registration). **

If you have already registered, please check your inbox for an email from Underline granting you access to ACL 2024 content.

Please register!

The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. More information will be announced soon.

Exploring and quantifying semantic relatedness is central to representing language and holds significant implications across various NLP tasks. While earlier NLP research primarily focused on semantic similarity, often within the English language context, we instead investigate the broader phenomenon of semantic relatedness. In this paper, we present \textit{SemRel}, a new semantic relatedness dataset collection annotated by native speakers across 13 languages: \textit{Afrikaans, Algerian Arabic, Amharic, English, Hausa, Hindi, Indonesian, Kinyarwanda, Marathi, Moroccan Arabic, Modern Standard Arabic, Spanish,} and \textit{Telugu}. These languages originate from five distinct language families and are predominantly spoken in Africa and Asia -- regions characterised by a relatively limited availability of NLP resources. Each instance in the SemRel datasets is a sentence pair associated with a score that represents the degree of semantic textual relatedness between the two sentences. The scores are obtained using a comparative annotation framework. We describe the data collection and annotation processes, challenges when building the datasets, baseline experiments, and their impact and utility in NLP.

SemRel2024: A Collection of Semantic Textual Relatedness Datasets for 13 Languages

The emergence of transformers has revolutionized natural language processing (NLP), as evidenced in various NLP tasks. While graph neural networks (GNNs) show recent promise in NLP, they are not standalone replacements for transformers. Rather, recent research explores combining transformers and GNNs. Existing GNN-based approaches rely on static graph construction methods requiring excessive text processing, and most of them are not scalable with the increasing document and word counts. We address these limitations by proposing a novel dynamic graph construction method for text documents based on vector visibility graphs (VVGs) generated from transformer output. Then, we introduce visibility pooler (VISPool), a scalable model architecture that seamlessly integrates VVG convolutional networks into transformer pipelines. We evaluate the proposed model on the General Language Understanding Evaluation (GLUE) benchmark datasets. VISPool outperforms the baselines with less trainable parameters, demonstrating the viability of the visibility-based graph construction method for enhancing transformers with GNNs.

VISPool: Enhancing Transformer Encoders with Vector Visibility Graph Neural Networks

Open-domain question answering (Open-QA) is a common task for evaluating large language models (LLMs). However, current Open-QA evaluations are criticized for the ambiguity in questions and the lack of semantic understanding in evaluators. Complex evaluators, powered by foundation models or LLMs and pertaining to semantic equivalence, still deviate from human judgments by a large margin. We propose to study the entailment relations of answers to identify more informative and more general system answers, offering a much closer evaluation to human judgment on both NaturalQuestions and TriviaQA while being learning-free. The entailment-based evaluation we propose allows the assignment of bonus or partial marks by quantifying the inference gap between answers, enabling a nuanced ranking of answer correctness that has higher AUC than current methods.

Accurate and Nuanced Open-QA Evaluation Through Textual Entailment

Multi-word expressions (MWEs) present unique challenges in natural language processing (NLP), particularly within the context of translation systems, due to their inherent scarcity, non-compositional nature, and other distinct lexical and morphosyntactic characteristics, issues that are exacerbated in low-resource settings.In this study, we elucidate and attempt to address these challenges by leveraging a substantial corpus of human-annotated Greek MWEs. To address the complexity of translating such phrases, we propose a novel method leveraging an available out-of-context lexicon.We assess the translation capabilities of current state-of-the-art systems on this task, employing both automated metrics and human evaluators.We find that by using our method when applicable, the performance of current systems can be significantly improved, however these models are still unable to produce translations comparable to those of a human speaker.

Dictionary-Aided Translation for Handling Multi-Word Expressions in Low-Resource Languages

As Natural Language Processing (NLP) systems become increasingly integrated into human social life, these technologies will need to increasingly rely on social intelligence. Although there are many valuable datasets that benchmark isolated dimensions of social intelligence, there does not yet exist any body of work to join these threads into a cohesive subfield in which researchers can quickly identify research gaps and future directions. Towards this goal, we build a Social AI Data Infrastructure, which consists of a comprehensive social AI taxonomy and a data library of 480 NLP datasets. Our infrastructure allows us to analyze existing dataset efforts, and also evaluate language models' performance in different social intelligence aspects. Our analyses demonstrate its utility in enabling a thorough understanding of current data landscape and providing a holistic perspective on potential directions for future dataset development. We show there is a need for multifaceted datasets, increased diversity in language and culture, more long-tailed social situations, and more interactive data in future social intelligence data efforts.

Social Intelligence Data Infrastructure: Structuring the Present and Navigating the Future

The prevalent approach for optimizing pre-trained language models in downstream tasks is fine-tuning. However, it is both time-consuming and memory-inefficient. In response, a more efficient method called Prefix Tuning, which insert learnable vectors into each Transformer layers, has been proposed and proven effective. Recent investigations reveal that prefix tokens carry context-specific information, prompting the hypothesis that enhancing their specialization can improve model performance. To address this, we propose Selective Prefix Tuning (SPT), integrating a selective mechanism inspired by selective self-attention. Additionally, we introduce Selective Loss (SL) to encourage diversity in prefix tokens. Extensive experiments validate the effectiveness of SPT in sentence and token classification tasks. We contribute insight into understanding the role of prefix in model adaptation.

Selective Prefix Tuning for Pre-trained Language Models

This paper surveys and organizes research works of medical dialog systems, which is an important yet challenging task. Although these systems have been surveyed in the medical community from an application perspective, a systematic review from a rigorous technical perspective has to date remained noticeably absent. As a result, an overview of the categories, methods, evaluation of medical dialogue systems remain limited and underspecified, hindering the further improvement of this area. To fill this gap, we investigate an initial pool of 325 papers from well-known computer science, natural language processing conferences and journals, and make an overview. Recently, large language models have shown strong model capacity on downstream tasks, which also reshape medical dialog systems' foundation.
Despite the alluring practical application value, current medical dialogue systems still suffer from problems. To this end, this paper lists grand challenges of medical dialog systems, especially of large language models.

Medical Dialogue System: A Survey of Categories, Methods, Evaluation and Challenges

The advent of large language models has experienced a remarkable improvement in the field of machine translation. However, machine translation is still vulnerable to critical meaning deviations, which may incur catastrophic issues in social or ethical contexts. In particular, existing critical error detection primarily focuses on identifying sentence-level errors, leaving the precise localization of such errors within the sentence unaddressed. In this paper, we introduce a new task, word-level critical error detection (WCED), to detect critical errors at a fine-grained level in machine translation sentences. The task aims to identify the parts of a machine translation that contain catastrophic meaning distortions. We hypothesize that the ability to determine errors at the sentence level will positively influence the detection of more granular errors. We propose a sentence-level error detection module to predict which words in a sentence have critical errors. Experimental results demonstrate that our method outperforms existing methodologies and LLM in En-De, Zh-En, En-Ru, and En-Ko. Our method is helpful for determining the fine-grained location of errors. We hope that such studies will improve the capacity to address critical errors adeptly.

Towards Precise Localization of Critical Errors in Machine Translation

Many studies have demonstrated that large language models (LLMs) can produce harmful responses, exposing users to unexpected risks. Previous studies have proposed comprehensive taxonomies of LLM risks, as well as corresponding prompts that can be used to examine LLM safety. However, the focus has been almost exclusively on English. We aim to broaden LLM safety research by introducing a dataset for the safety evaluation of Chinese LLMs, and extending it to better identify false negative and false positive examples in terms of risky prompt rejections. We further present a set of fine-grained safety assessment criteria for each risk type, facilitating both manual annotation and automatic evaluation in terms of LLM response harmfulness. Our experiments over five LLMs show that region-specific risks are the prevalent risk type. Warning: this paper contains example data that may be offensive, harmful, or biased. Our data is available at https://github.com/Libr-AI/do-not-answer.

A Chinese Dataset for Evaluating the Safeguards in Large Language Models

Large Language Models (LLMs) are widely used to evaluate natural language generation tasks as automated metrics.
However, the likelihood, a measure of LLM's plausibility for a sentence, can vary due to superficial differences in sentences, such as word order and sentence structure.
It is therefore possible that there might be a likelihood bias if LLMs are used for evaluation: they might overrate sentences with higher likelihoods while underrating those with lower likelihoods.
In this paper, we investigate the presence and impact of likelihood bias in LLM-based evaluators.
We also propose a method to mitigate the likelihood bias.
Our method utilizes highly biased instances as few-shot examples for in-context learning.
Our experiments in evaluating the data-to-text and grammatical error correction tasks reveal that several LLMs we test display a likelihood bias.
Furthermore, our proposed method successfully mitigates this bias, also improving evaluation performance (in terms of correlation of models with human scores) significantly.

Premium content

Ranking Large Language Models without Ground Truth

Downloads

Next from ACL 2024

SemRel2024: A Collection of Semantic Textual Relatedness Datasets for 13 Languages

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES