The rapid advancement of Large Language Models (LLMs) has improved text understanding and generation but poses challenges in computational resources. This study proposes a curriculum learning-inspired, data-centric training strategy that begins with simpler tasks and progresses to more complex ones, using criteria such as prompt length, attention scores, and loss values to structure the training data. Experiments with Mistral-7B (Jiang et al., 2023) and Gemma-7B (Team et al., 2024) models demonstrate that curriculum learning slightly improves performance compared to traditional random data shuffling. Notably, we observed that sorting data based on our proposed attention criteria generally led to better performance. This approach offers a sustainable method to enhance LLM performance without increasing model size or dataset volume, addressing scalability challenges in LLM training.

For language model classification, would you prefer having only one workable class or having every class working? The latter makes more practical uses. Especially for large language models (LLMs), the fact that they achieve a fair overall accuracy by in-context learning (ICL) obscures a large difference in individual class accuracies. In this work, we uncover and tackle language models' imbalance in per-class prediction accuracy by reconceptualizing it as the Contextual Oddity Bias (COBias), and we are the first to engage nonlinear integer programming (NIP) to debias it. Briefly, COBias refers to the difference in accuracy by a class A compared to its ''odd'' class, which holds the majority wrong predictions of class A. With the COBias metric, we reveal that LLMs of varied scales and families exhibit large per-class accuracy differences. Then we propose Debiasing as Nonlinear Integer Programming (DNIP) to correct ICL per-class probabilities for lower bias and higher overall accuracy. Our optimization objective is directly based on the evaluation scores by COBias and accuracy metrics, solved by simulated annealing. Evaluations on three LLMs across seven NLP classification tasks show that DNIP simultaneously achieves significant COBias reduction ($-27\%$) and accuracy improvement ($+12\%$) over the conventional ICL approach, suggesting that modeling pairwise class accuracy differences is a direction in pushing forward more accurate, more reliable LLM predictions.

Word embeddings are fixed-length, dense and distributed word representations that are used in natural language processing (NLP) applications. There are basically two types of word embedding models which are non-contextual (static) models and contextual models. The former method generates a single embedding for a word regardless of its context, while the latter method produces distinct embeddings for a word based on the specific contexts in which it appears. There are plenty of works that compare contextual and non-contextual embedding models within their respective groups in different languages. However, the number of studies that compare the models in these two groups with each other is very few and there is no such study in Turkish. This process necessitates converting contextual embeddings into static embeddings. In this paper, we compare and evaluate the performance of several contextual and non-contextual models in both intrinsic and extrinsic evaluation settings for Turkish. We make a fine-grained comparison by analyzing the syntactic and semantic capabilities of the models separately. The results of the analyses provide insights about the suitability of different embedding models in different types of NLP tasks. We also build a Turkish word embedding repository comprising the embedding models used in this work, which may serve as a valuable resource for researchers and practitioners in the field of Turkish NLP. We make the word embeddings, scripts, and evaluation datasets publicly available.

When large language models are aligned via supervised fine-tuning, they may encounter new factual information that was not acquired through pre-training. It is often conjectured that this can teach the model the behavior of hallucinating factually incorrect responses, as the model is trained to generate facts that are not grounded in its pre-existing knowledge. In this work, we study the impact of such exposure to new knowledge on the capability of the fine-tuned model to utilize its pre-existing knowledge. To this end, we design a controlled setup, focused on closed-book QA, where we vary the proportion of the fine-tuning examples that introduce new knowledge. We demonstrate that large language models struggle to acquire new factual knowledge through fine-tuning, as fine-tuning examples that introduce new knowledge are learned significantly slower than those consistent with the model's knowledge. However, we also find that as the examples with new knowledge are eventually learned, they linearly increase the model's tendency to hallucinate. Taken together, our results highlight the risk in introducing new factual knowledge through fine-tuning, and support the view that large language models mostly acquire factual knowledge through pre-training, whereas fine-tuning teaches them to use it more efficiently.

Persona-driven role-playing (PRP) aims to build AI characters that can respond to user queries by faithfully sticking with all persona statements. Unfortunately, existing faithfulness criteria for PRP are limited to coarse-grained LLM-based scoring without a clear definition or formulation. This paper presents a pioneering exploration to quantify PRP faithfulness as a fine-grained and explainable criterion, which also serves as a reliable reference for optimization. Our criterion first discriminates persona statements into active and passive constraints by identifying the query-statement relevance. Then, we incorporate all constraints following the principle that the AI character's response should be (a) entailed by active (relevant) constraints and (b) not contradicted by passive (irrelevant) constraints. We translate this principle mathematically into a novel Active-Passive-Constraint (APC) score, a constraint-wise sum of natural language inference (NLI) scores weighted by relevance scores. In practice, we build the APC scoring system by symbolically distilling small discriminators from GPT-4 for efficiency. We validate the quality of the APC score against human evaluation based on example personas with tens of statements, and the results show a high correlation. We further leverage it as a reward system in direct preference optimization (DPO) for better AI characters. Our experiments offer a fine-grained and explainable comparison between existing PRP techniques, revealing their advantages and limitations. We further find APC-based DPO to be one of the most competitive techniques for sticking with all constraints and can be well incorporated with other techniques. We then extend the scale of the experiments to real persons with hundreds of statements and reach a consistent conclusion.

The remarkable progress of Multi-modal Large Language Models (MLLMs) has attracted significant attention due to their superior performance in visual contexts. However, their capabilities in turning visual figure to executable code, have not been evaluated thoroughly. To address this, we introduce Plot2Code, a comprehensive visual coding benchmark designed for a fair and in-depth assessment of MLLMs. We carefully collect 132 manually selected high-quality matplotlib plots across six plot types from publicly available matplotlib galleries. For each plot, we carefully offer its source code, and an descriptive instruction summarized by GPT-4. This approach enables Plot2Code to extensively evaluate MLLMs' code capabilities across various input modalities. Furthermore, we propose three automatic evaluation metrics, including code pass rate, text-match ratio, and GPT-4V overall rating, for a fine-grained assessment of the output code and rendered images. Instead of simply judging pass or fail, we employ GPT-4V to make an overall judgement between the generated and reference images, which has been shown to be consistent with human evaluation. The evaluation results, which include analyses of 14 MLLMs such as the proprietary GPT-4V, Gemini-Pro, and the open-sourced Mini-Gemini, highlight the substantial challenges presented by Plot2Code. With Plot2Code, we reveal that most existing MLLMs struggle with visual coding for text-dense plots, heavily relying on textual instruction. We hope that the evaluation results from Plot2Code on visual coding will guide the future development of MLLMs. All data involved with Plot2Code are available at https://huggingface.co/datasets/TencentARC/Plot2Code.

Personalized news recommendations are essential for online news platforms to assist users in discovering news articles that match their interests from a vast amount of online content. Appropriately encoded content features, such as text, categories, and images, are essential for recommendations. Among these features, news categories, such as tv-golden-globe, finance-real-estate, and news-politics, play an important role in understanding news content, inspiring us to enhance the categories' descriptions. In this paper, we propose a novel method that automatically generates informative category descriptions using a large language model (LLM) without manual effort or domain-specific knowledge and incorporates them into recommendation models as additional information. In our comprehensive experimental evaluations using the MIND dataset, our method successfully achieved 5.8% improvement at most in AUC compared with baseline approaches without the LLM's generated category descriptions for the state-of-the-art content-based recommendation models including NAML, NRMS, and NPA. These results validate the effectiveness of our approach. The code is available at https://github.com/yamanalab/gpt-augmented-news-recommendation.

This paper introduces v0.5 of the AI Safety Benchmark, which has been created by the MLCommons AI Safety Working Group. The AI Safety Benchmark has been designed to assess the safety risks of AI systems that use chat-tuned language models. We introduce a principled approach to specifying and constructing the benchmark, which for v0.5 covers only a single use case (an adult chatting to a general-purpose assistant in English), and a limited set of personas (i.e., typical users, malicious users, and vulnerable users). We created a new taxonomy of 13 hazard categories, of which 7 have tests in the v0.5 benchmark. We plan to release version 1.0 of the AI Safety Benchmark by the end of 2024. The v1.0 benchmark will provide meaningful insights into the safety of AI systems. However, the v0.5 benchmark should not be used to assess the safety of AI systems. We have sought to fully document the limitations, flaws, and challenges of v0.5. This release of v0.5 of the AI Safety Benchmark includes (1) a principled approach to specifying and constructing the benchmark, which comprises use cases, types of systems under test (SUTs), language and context, personas, tests, and test items; (2) a taxonomy of 13 hazard categories with definitions and subcategories; (3) tests for seven of the hazard categories, each comprising a unique set of test items, i.e., prompts. There are 43,090 test items in total, which we created with templates; (4) a grading system for AI systems against the benchmark; (5) an openly available platform, and downloadable tool, called ModelBench that can be used to evaluate the safety of AI systems on the benchmark; (6) an example evaluation report which benchmarks the performance of over a dozen openly available chat-tuned language models; (7) a test specification for the benchmark.

Existing text-to-image models struggle to follow complex text prompts, raising the need for extra grounding inputs for better controllability. In this work, we propose to decompose a scene into visual primitives - denoted as dense blob representations - that contain fine-grained details of the scene while being modular, human-interpretable, and easy-to-construct. Based on blob representations, we develop a blob-grounded text-to-image diffusion model, termed BlobGEN, for compositional generation. Particularly, we introduce a new masked cross-attention module to disentangle the fusion between blob representations and visual features. To leverage the compositionality of large language models (LLMs), we introduce a new in-context learning approach to generate blob representations from text prompts. Our extensive experiments show that BlobGEN achieves superior zero-shot generation quality and better layout-guided controllability on MS-COCO. When augmented by LLMs, our method exhibits superior numerical and spatial correctness on compositional image generation benchmarks.

Many natural language processing systems operate over tokenizations of text to address the open-vocabulary problem. In this paper, we give and analyze an algorithm for the efficient construction of deterministic finite automata designed to operate directly on tokenizations produced by the popular byte pair encoding technique. This makes it possible to apply many existing techniques and algorithms to the tokenized case, such as pattern matching, equivalence checking of tokenization dictionaries, and composing tokenized languages in various ways.