Thailand

The natural language processing field has been evolving around language models for the past few years, from the usage of n-gram language models for re-ranking, to transfer learning with encoder-only (BERT-like) language models, and finally to large language models (LLMs) as general solvers. LLMs are dominated by the decoder-only type, and they are popular for their efficacy in numerous tasks. LLMs are regarded as having strong comprehension abilities and strong capabilities to solve new unseen tasks. As such, people may quickly assume that decoder-only LLMs always perform better than the encoder-only ones, especially for understanding word meaning. In this paper, we demonstrate that decoder-only LLMs perform worse on word meaning comprehension than an encoder-only language model that has vastly fewer parameters.

ACL 2024

Are Decoder-Only Language Models Better than Encoder-Only Language Models in Understanding Word Meaning?

word-in-context

word sense disambiguation

large language models

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.

Since the natural language processing (NLP) community started to make large language models (LLMs) act as a critic to evaluate the quality of generated texts, most of the existing works train a critique generation model on the evaluation data labeled by GPT-4's direct prompting. We observe that these models lack the ability to generate informative critiques in both pointwise grading and pairwise comparison especially without references. As a result, their generated critiques cannot provide fine-grained distinguishability on generated texts, causing unsatisfactory evaluation performance. In this paper, we propose a simple yet effective method called Eval-Instruct, which can first acquire pointwise grading critiques with pseudo references and then revise these critiques via multi-path prompting to obtain informative evaluation data in different tasks and settings, including pointwise grading and pairwise comparison with / without references. After fine-tuning on these data, the resulting model CritiqueLLM is empirically shown to outperform ChatGPT and all the open-source baselines and even achieve comparable evaluation performance to GPT-4 in system-level correlations of pointwise grading. We also demonstrate that our generated critiques can act as scalable feedback to further improve the generation quality of strong LLMs like ChatGPT.

CritiqueLLM: Towards an Informative Critique Generation Model for Evaluation of Large Language Model Generation

Instruction tuning enhances the instruction following ability of large language models by finetuning with supervised instruction data. Previous work proposes in-context instruction tuning (ICIT) where specific positive or negative examples are incorporated into the prompt for better performance. In this work, we propose PACIT, a simple and effective in-context instruction tuning method, inspired by the pedagogical concept of desirable difficulty. The PACIT method unlocks the power of examples by encouraging the model to actively learn to grasp the distinctions between the positive and negative examples instead of merely reading. The model is expected to first verify the correctness of the provided example according to the task description, which is then set as the condition for generating a better response to the task instance. Our extensive experiments prove the effectiveness of PACIT, outperforming ICIT baseline on both in-domain and out-domain tasks up to 9.16 and 3.14 average ROUGE-L scores, respectively. Moreover, PACIT can notably enhance the performance of instruction tuning even when all positive and negative examples are generated with a self-instruct method.

PACIT: Unlocking the Power of Examples for Better In-Context Instruction Tuning

Deploying large language models (LLMs) to real scenarios for domain-specific question answering (QA) is a key thrust for LLM applications, which poses numerous challenges, especially in ensuring that responses are both accommodating to user requirements and appropriately leveraging domain-specific knowledge bases. They are the two major difficulties for LLM application as vanilla fine-tuning falls short of addressing. Combining these requirements, we conceive of them as the requirement for the model's preference to be harmoniously aligned with humans'. Thus, we introduce Knowledgeable Preference AlignmenT (KnowPAT), which constructs two kinds of preference sets to tackle the two issues. Besides, we design a new alignment objective to align the LLM preference with different human preferences uniformly, aiming to optimize LLM performance in real-world, domain-specific QA settings. Adequate experiments and comprehensive comparisons with 15 baseline methods illustrate that our KnowPAT is a superior pipeline for real-scenario domain-specific QA with LLMs.

Knowledgeable Preference Alignment for LLMs in Domain-specific Question Answering

Large language models demonstrate a remarkable capability for learning to solve new tasks from a few examples.
The $\textit{prompt template}$, or the way the input examples are formatted to obtain the prompt, is an important yet often overlooked aspect of in-context learning.
In this work, we conduct a comprehensive study of the template format's influence on the in-context learning performance.
We evaluate the impact of the prompt template across 21 models (from 770M to 70B parameters) and 4 standard classification datasets. 
We show that a poor choice of the template can reduce the performance of the strongest models and inference methods to a random guess level.
More importantly, the best templates do not transfer between different setups and even between models of the same family.
Our findings show that the currently prevalent approach to evaluation, which ignores template selection, may give misleading results due to different templates in different works.
As a first step towards mitigating this issue, we propose $\textit{Template Ensembles}$ that aggregate model predictions across several templates.
This simple test-time augmentation boosts average performance while being robust to the choice of random set of templates.

Mind Your Format: Towards Consistent Evaluation of In-Context Learning Improvements

This paper introduces the ColorSwap dataset, designed to assess and improve the proficiency of multimodal models in matching objects with their colors. The dataset is comprised of 2,000 unique image-caption pairs, grouped into 1,000 examples. Each example includes a caption-image pair, along with a "color-swapped" pair. We follow the Winoground schema: the two captions in an example have the same words, but the color words have been rearranged to modify different objects. The dataset was created through a novel blend of automated caption and image generation with humans in the loop. We evaluate image-text matching (ITM) and visual language models (VLMs) and find that even the latest ones are still not robust at this task. GPT-4V and LLaVA score 72% and 42% on our main VLM metric, although they may improve with more advanced prompting techniques. On the main ITM metric, contrastive models such as CLIP and SigLIP perform close to chance (at 12% and 30%, respectively), although the non-contrastive BLIP ITM model is stronger (87%). We also find that finetuning on fewer than 2,000 examples yields significant performance gains on this out-of-distribution word-order understanding task.

ColorSwap: A Color and Word Order Dataset for Multimodal Evaluation

Text-to-image synthesis for the Chinese language poses unique challenges due to its large vocabulary size, and intricate character relationships. While existing diffusion models have shown promise in generating images from textual descriptions, they often neglect domain-specific contexts and lack robustness in handling the Chinese language. This paper introduces PAI-Diffusion, a comprehensive framework that addresses these limitations. PAI-Diffusion incorporates both general and domain-specific Chinese diffusion models, enabling the generation of contextually relevant images. It explores the potential of using LoRA and ControlNet for fine-grained image style transfer and image editing, empowering users with enhanced control over image generation. Moreover, PAI-Diffusion seamlessly integrates with Alibaba Cloud's Platform for AI, providing accessible and scalable solutions. All the Chinese diffusion model checkpoints, LoRAs, and ControlNets, including domain-specific ones, are publicly available. A user-friendly Chinese WebUI and the diffusers-api elastic inference toolkit, also open-sourced, further facilitate the easy deployment of PAI-Diffusion models in various local and cloud environments, making it a valuable resource for Chinese text-to-image synthesis.

PAI-Diffusion: Constructing and Serving a Family of Open Chinese Diffusion Models for Text-to-image Synthesis on the Cloud

Instruction tuning significantly enhances the performance of large language models (LLMs) across various tasks. However, the procedure to optimizing the mixing of instruction datasets for LLM fine-tuning is still poorly understood. This study categorizes instructions into three primary types: NLP downstream tasks, coding, and general chat. We explore the effects of instruction tuning on different combinations of datasets on LLM performance, and find that certain instruction types are more advantageous for specific applications but can negatively impact other areas. This work provides insights into instruction mixtures, laying the foundations for future research.

Demystifying Instruction Mixing for Fine-tuning Large Language Models

Researchers in political and social sciences often rely on classification models to analyze trends in information consumption by examining browsing histories of millions of webpages. Automated scalable methods are necessary due to the impracticality of manual labeling. In this paper, we model the detection of policy-related content as binary classification task and compare the accuracy of fine-tuned pre-trained encoder models against in-context learning strategies. Using only a few hundred annotated data points per topic, we detect content related to three German policies in a database of scraped webpages. We compare multilingual and monolingual models, as well as zero and few-shot approaches, and investigate the impact of negative sampling strategies and the combination of URL and content-based features. Our research shows that a small sample of annotated data is sufficient to train an effective classifier. Fine-tuning encoder-based models yields better results than in-context learning. Classifiers using both URL and content-based features perform best, while URL alone provides adequate accuracy when content is unavailable.

Assessing In-context Learning and Fine-tuning for Topic Classification of German Web Data

The processing of long sequences with models remains a subject in its own right, including automatic summary, despite recent improvements. In this work, we present experiments on the automatic summarization of scientific articles using BART models, taking into account textual information coming from distinct passages from the long texts to be summarized. We demonstrate that taking into account document structure improves the performance of state-of-the-art models and approaches the performance of LongFormer on English.

Does the structure of textual content have an impact on language models for automatic summarization?

This paper investigates the robustness of Large Language Models (LLMs) against Out-Of-Distribution (OOD) data within the context of sentiment analysis. Traditional fine-tuning approaches often fail to generalize effectively across different data distributions, limiting the practical deployment of LLMs in dynamic real-world scenarios. To address this challenge, we introduce a novel method called "Semantic Rewriting," which leverages the inherent flexibility of LLMs to align both in-distribution (ID) and OOD data with the LLMs distributions. By semantically transforming sentences to minimize linguistic discrepancies, our approach helps to standardize features across datasets, thus enhancing model robustness. We conduct extensive experiments with several benchmark datasets and LLMs to validate the efficacy of our method. The results demonstrate that Semantic Rewriting significantly improves the performance of models on OOD tasks, outperforming traditional methods in both robustness and generalization capabilities. Our findings suggest that Semantic Rewriting is a promising technique for developing more reliable and versatile NLP systems capable of performing robustly across diverse operational environments.

Premium content

Are Decoder-Only Language Models Better than Encoder-Only Language Models in Understanding Word Meaning?

Downloads

Next from ACL 2024

CritiqueLLM: Towards an Informative Critique Generation Model for Evaluation of Large Language Model Generation

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES