Diversity of Transformer Layers: One Aspect of Parameter Scaling Laws

Transformers deliver outstanding performance across a wide range of tasks and are now a dominant backbone architecture for large language models (LLMs). Their task-solving performance is improved by increasing parameter size, as shown in the recent studies on parameter scaling laws. Although recent mechanistic-interpretability studies have deepened our understanding of the internal behavior of Transformers by analyzing their residual stream, the relationship between these internal mechanisms and the parameter scaling laws remains unclear. To bridge this gap, we focus on layers and their size, which mainly decide the parameter size of Transformers. For this purpose, we first theoretically investigate the layers within the residual stream through a bias-diversity decomposition. The decomposition separates (i) bias, the error of each layer's output from the ground truth, and (ii) diversity, which indicates how much the outputs of each layer differ from each other. Analyzing Transformers under this theory reveals that performance improves when individual layers make predictions close to the correct answer and remain mutually diverse. We show that diversity becomes especially critical when individual layers' outputs are far from the ground truth. Finally, we introduce an information-theoretic diversity and show our main findings that adding layers enhances performance only when those layers behave differently, i.e., are diverse. We also reveal the performance gains from increasing the number of layers exhibit submodularity: marginal improvements diminish as additional layers increase, mirroring the logarithmic convergence predicted by the parameter scaling laws. Experiments on multiple semantic-understanding tasks with various LLMs empirically confirm the theoretical properties derived in this study.

Do LLMs Need to Think in One Language? Correlation between Latent Language and Task Performance

Large Language Models (LLMs) are known to process information using a proficient internal language consistently, referred to as latent language, which may differ from the input or output languages. However, how the discrepancy between the latent language and the input and output language affects downstream task performance remains largely unexplored. While many studies research the latent language of LLMs, few address its importance in influencing task performance. In our study, we hypothesize that thinking in latent language consistently enhances downstream task performance. To validate this, our work varies the input prompt languages across multiple downstream tasks and analyzes the correlation between consistency in latent language and task performance. We create datasets consisting of questions from diverse domains such as translation and geo-culture, which are influenced by the choice of latent language. Experimental results across multiple LLMs on translation and geo-culture tasks, which are sensitive to the choice of language, indicate that maintaining consistency in latent language is not always necessary for optimal downstream task performance. This is because these models adapt their internal representations near the final layers to match the target language, reducing the impact of consistency on overall performance.

gec-metrics: A Unified Library for Grammatical Error Correction Evaluation

We introduce gec-metrics, a library for using and developing grammatical error correction (GEC) evaluation metrics through a unified interface. Our library enables fair system comparisons by ensuring that everyone conducts evaluations using a consistent implementation. Moreover, it is designed with a strong focus on API usage, making it highly extensible. It also includes meta-evaluation functionalities and provides analysis and visualization scripts, contributing to developing GEC evaluation metrics. Our code is released under the MIT license and is also distributed as an installable package. The video is available on YouTube.

Diagnosing Vision Language Models' Perception by Leveraging Human Methods for Color Vision Deficiencies

Large-scale Vision Language Models (LVLMs) are increasingly being applied to a wide range of real-world multimodal applications, involving complex visual and linguistic reasoning. As these models become more integrated into practical use, they are expected to handle complex aspects of human interaction. Among these, color perception is a fundamental yet highly variable aspect of visual understanding. It differs across individuals due to biological factors such as Color Vision Deficiencies (CVDs), as well as differences in culture and language. Despite its importance, perceptual diversity has received limited attention. In our study, we evaluate LVLMs' ability to account for individual level perceptual variation using the Ishihara Test, a widely used method for detecting CVDs. Our results show that LVLMs can explain CVDs in natural language, but they cannot simulate how people with CVDs perceive color in image based tasks. These findings highlight the need for multimodal systems that can account for color perceptual diversity and support broader discussions on perceptual inclusiveness and fairness in multimodal AI.

IterKey: Iterative Keyword Generation with LLMs for Enhanced Retrieval Augmented Generation

Retrieval-Augmented Generation (RAG) has emerged as a way to complement the in-context knowledge of Large Language Models (LLMs) by integrating external documents. However, real-world applications demand not only accuracy but also interpretability. While dense retrieval methods provide high accuracy, they lack interpretability; conversely, sparse retrieval methods offer transparency but often fail to capture the full intent of queries due to their reliance on keyword matching. To address these issues, we introduce IterKey, an LLM-driven iterative keyword generation framework that enhances RAG via sparse retrieval. IterKey consists of three LLM-driven stages: generating keywords for retrieval, generating answers based on retrieved documents, and validating the answers. If validation fails, the process iteratively repeats with refined keywords. Across four QA tasks, experimental results show that IterKey achieves 5% to 20% accuracy improvements over BM25-based RAG and simple baselines. Its performance is comparable to dense retrieval-based RAG and prior iterative query refinement methods using dense models. In summary, IterKey is a novel BM25-based approach leveraging LLMs to iteratively refine RAG, effectively balancing accuracy with interpretability.

TextTIGER: Text-based Intelligent Generation with Entity Prompt Refinement for Text-to-Image Generation

Generating images from prompts containing specific entities requires models to retain as much entity-specific knowledge as possible. However, fully memorizing such knowledge is impractical due to the vast number of entities and their continuous emergence. To address this, we propose Text-based Intelligent Generation with Entity prompt Refinement (TextTIGER), which augments knowledge on entities included in the prompts and then summarizes the augmented descriptions using Large Language Models (LLMs) to mitigate performance degradation from longer inputs. To evaluate our method, we introduce WiT-Cub (WiT with Captions and Uncomplicated Background-explanations), a dataset comprising captions, images, and an entity list. Experiments on four image generation models and five LLMs show that TextTIGER improves image generation performance in standard metrics (IS, FID, and CLIPScore) compared to caption-only prompts. Additionally, multiple annotators' evaluation confirms that the summarized descriptions are more informative, validating LLMs' ability to generate concise yet rich descriptions. These findings demonstrate that refining prompts with augmented and summarized entity-related descriptions enhances image generation capabilities. The code and dataset will be available upon acceptance.

Long-Tail Crisis in Nearest Neighbor Language Models

The $k$-nearest-neighbor language model ($k$NN-LM), one of the retrieval-augmented language models, improves the perplexity for given text by directly accessing a large datastore built from any text data during inference. A widely held hypothesis for the success of $k$NN-LM is that its explicit memory, i.e., the datastore, enhances predictions for long-tail phenomena. However, prior works have primarily shown its ability to retrieve long-tail contexts, leaving the model's performance remain underexplored in estimating the probabilities of long-tail target tokens during inference. In this paper, we investigate the behavior of $k$NN-LM on low-frequency tokens, examining prediction probability, retrieval accuracy, token distribution in the datastore, and approximation error of the product quantization. Our experimental results reveal that $k$NN-LM does not improve prediction performance for low-frequency tokens but mainly benefits high-frequency tokens regardless of long-tail contexts in the datastore.

Rethinking Evaluation Metrics for Grammatical Error Correction: Why Use a Different Evaluation Process than Human?

One of the goals of automatic evaluation metrics in grammatical error correction (GEC) is to rank GEC systems such that it matches human preferences. However, current automatic evaluations are based on procedures that diverge from human evaluation. Specifically, human evaluation derives rankings by aggregating sentence-level relative evaluation results, e.g., pairwise comparisons, using a rating algorithm, whereas automatic evaluation averages sentence-level absolute scores to obtain corpus-level scores, which are then sorted to determine rankings. In this study, we propose an aggregation method for existing automatic evaluation metrics which aligns with human evaluation methods to bridge this gap. We conducted experiments using various metrics, including edit-based metrics, $n$-gram based metrics, and sentence-level metrics, and show that resolving the gap improves results for the most of metrics on the SEEDA benchmark. We also found that even BERT-based metrics sometimes outperform the metrics of GPT-4. We publish our unified implementation of the metrics and meta-evaluations.

Tonguescape: Exploring Language Models Understanding of Vowel Articulation

Vowels are primarily characterized by tongue position. Humans have discovered these features of vowel articulation through their own experience and explicit objective observation such as using MRI. With this knowledge and our experience, we can explain and understand the relationship between tongue positions and vowels, and this knowledge is helpful for language learners to learn pronunciation. Since language models (LMs) are trained on a large amount of data that includes linguistic and medical fields, our preliminary studies indicate that an LM is able to explain the pronunciation mechanisms of vowels. However, it is unclear whether multi-modal LMs, such as vision LMs, align textual information with visual information. One question arises: do LMs associate real tongue positions with vowel articulation? In this study, we created video and image datasets from the existing real-time MRI dataset and investigated whether LMs can understand vowel articulation based on tongue positions using vision-based information. Our findings suggest that LMs exhibit potential for understanding vowels and tongue positions when reference examples are provided while they have difficulties without them. Our code for dataset building is available on GitHub.

Measuring the Robustness of Reference-Free Dialogue Evaluation Systems

Advancements in dialogue systems powered by large language models (LLMs) have outpaced the development of reliable evaluation metrics, particularly for diverse and creative responses. We present a benchmark for evaluating the robustness of reference-free dialogue metrics against four categories of adversarial attacks: speaker tag prefixes, static responses, ungrammatical responses, and repeated conversational context. We analyze metrics such as DialogRPT, UniEval, and PromptEval -- a prompt-based method leveraging LLMs -- across grounded and ungrounded datasets. By examining both their correlation with human judgment and susceptibility to adversarial attacks, we find that these two axes are not always aligned; metrics that appear to be equivalent when judged by traditional benchmarks may, in fact, vary in their scores of adversarial responses. These findings motivate the development of nuanced evaluation frameworks to address real-world dialogue challenges.