LLM-as-a-Judge is a framework that uses an LLM (large language model) to evaluate the quality of natural language text - typically text that is also generated by an LLM. This framework holds great promise due to its relative low-cost, ease of use, and strong correlations with human stylistic preferences. However, LLM Judges have been shown to exhibit biases that can distort their judgments. We evaluate how well LLM Judges can grade whether a given response to a conversational question is correct, an ability crucial to soundly estimating the overall response quality. To do so, we create and publicly release a human-annotated dataset with labels of correctness for 1,200 LLM responses. We source questions from a combination of existing datasets and a novel, challenging benchmark (BFF-Bench) created for this analysis. We demonstrate a strong connection between an LLM's ability to correctly answer a question and grade responses to that question. Although aggregate level statistics might imply a judge has high agreement with human annotators, it will struggle on the subset of questions it could not answer. To address this issue, we recommend a simple solution: provide the judge with a correct, human-written reference answer. We perform an in-depth analysis on how reference quality can affect the performance of an LLM Judge. We show that providing a weaker judge (e.g. Qwen 2.5 7B) with higher quality references reaches better agreement with human annotators than a stronger judge (e.g. GPT-4o) with synthetic references.

Tokenization, a crucial initial step in natural language processing, is often assumed to benefit from larger training datasets. This paper investigates the impact of tokenizer training data sizes ranging from 1GB to 900GB. Our findings reveal diminishing returns as the data size increases, highlighting a practical limit on how much further scaling the training data can improve tokenization quality. We analyze this phenomenon and attribute the saturation effect to the constraints imposed by the pre-tokenization stage of tokenization. These results offer valuable insights for optimizing the tokenization process and highlight potential avenues for future research in tokenization algorithms.

Some information is factual (e.g., "Paris is in France"), whereas other information is probabilistic (e.g., "the coin flip will be a [Heads/T ails]."). We believe that good Language Models (LMs) should understand and reflect this nuance. Our work investigates this by testing if LMs' output probabilities are calibrated to their textual contexts. We define model "calibration" as the degree to which the output probabilities of candidate tokens are aligned with the relative likelihood that should be inferred from the given context. For example, if the context concerns two equally likely options (e.g., heads or tails for a fair coin), the output probabilities should reflect this. Likewise, context that concerns non-uniformly likely events (e.g., rolling a six with a die) should also be appropriately captured with proportionate output probabilities. We find that even in simple settings the best LMs (1) are poorly calibrated, and (2) have systematic biases (e.g., preferred colors and sensitivities to word orderings). For example, gpt-4o-mini often picks the first of two options presented in the prompt regardless of the options' implied likelihood, whereas Llama-3.1-8B Our other consistent finding is mode-collapse: Instruction-tuned models often over-allocate probability mass on a single option. These systematic biases introduce non-intuitive model behavior, making models harder for users to understand. We investigate the extent to which language model (LM) output probabilities are calibrated to the numeric content of their contexts. Figure 1: Models produce un-calibrated results. Inputting Examples 1 and 2 to gpt-4o different, uncalibrated behaviors arise in the model probabilities.

The financial domain frequently deals with large numbers of long documents that are essential for daily operations. Significant effort is put towards automating financial data analysis. However, a persistent challenge, not limited to the finance domain, is the scarcity of datasets that accurately reflect real-world tasks for model evaluation. Existing datasets are often constrained by size, context, or relevance to practical applications. Moreover, LLMs are currently trained on trillions of tokens of text, limiting access to novel data or documents that models have not encountered during training for unbiased evaluation. We propose SEC-QA, a continuous dataset generation framework with two key features: 1) the semi-automatic generation of Question-Answer (QA) pairs spanning multiple long context financial documents, which better represent real-world financial scenarios; 2) the ability to continually refresh the dataset using the most recent public document collections, not yet ingested by LLMs. Our experiments show that current retrieval augmented generation methods systematically fail to answer these challenging multi-document questions. In response, we introduce a QA system based on program-of-thought that improves the ability to perform complex information retrieval and quantitative reasoning pipelines, thereby increasing QA accuracy.

While subword tokenizers such as BPE and WordPiece are typically used to build vocabularies for NLP models, the method of decoding text into a sequence of tokens from these vocabularies is often left unspecified, or ill-suited to the method in which they were constructed. We provide a controlled analysis of seven tokenizer inference methods across four different algorithms and three vocabulary sizes, performed on a novel intrinsic evaluation suite we curated for English, combining measures rooted in morphology, cognition, and information theory. We show that for the most commonly used tokenizers, greedy inference performs surprisingly well; and that SaGe, a recently-introduced contextually-informed tokenizer, outperforms all others on morphological alignment.

Tokenization is a foundational step in Natural Language Processing (NLP) tasks, bridging raw text and language models. Existing tokenization approaches like Byte-Pair Encoding (BPE) originate from the field of data compression, and it has been suggested that the effectiveness of BPE stems from its ability to condense text into a relatively small number of tokens. We test the hypothesis that fewer tokens lead to better downstream performance by introducing PathPiece, a new tokenizer that segments a document's text into the minimum number of tokens for a given vocabulary. Through extensive experimentation we find this hypothesis not to be the case, casting doubt on the understanding of the reasons for effective tokenization. To examine which other factors play a role, we evaluate design decisions across all three phases of tokenization: pre-tokenization, vocabulary construction, and segmentation, offering new insights into the design of effective tokenizers. Specifically, we illustrate the importance of pre-tokenization and the benefits of using BPE to initialize vocabulary construction. We train 64 language models with varying tokenization, ranging in size from 350M to 2.4B parameters, all of which are made publicly available.

Research in quantitative reasoning within the financial domain indeed necessitates the use of realistic tasks and data, primarily because of the significant impact of decisions made in business and finance. Financial professionals often interact with documents hundreds of pages long, but most research datasets drastically reduce this context length. To address this, we introduce a long-document financial QA task. We augment 7,621 questions from the existing FinQA dataset with full-document context, extending the average context length for each question from under 700 words in FinQA to 123k words in DocFinQA. We conduct extensive experiments of retrieval-based QA pipelines and long-context language models on the augmented data. Our results show that DocFinQA provides challenges for even the strongest, state-of-the-art systems.

As large language models (LLMs) impact a growing number of complex domains, it is becoming increasingly important to have fair, accurate, and rigorous evaluation benchmarks. Evaluating the reasoning skills required for business and financial NLP stands out as a particularly difficult challenge. We introduce BizBench, a new benchmark for evaluating models' ability to reason about realistic financial problems. BizBench comprises 8 quantitative reasoning tasks. Notably, BizBench targets the complex task of question-answering (QA) for structured and unstructured financial data via program synthesis (i.e., code generation). We introduce three diverse financially-themed code-generation tasks from newly collected and augmented QA data. Additionally, we isolate distinct financial reasoning capabilities required to solve these QA tasks: reading comprehension of financial text and tables, which is required to extract correct intermediate values; and understanding domain knowledge (e.g., financial formulas) needed to calculate complex solutions. Collectively, these tasks evaluate a model's financial background knowledge, ability to extract numeric entities from financial documents, and capacity to solve problems with code. We conduct an in-depth evaluation of open-source and commercial LLMs, illustrating that BizBench is a challenging benchmark for quantitative reasoning in the finance and business domain.

Document layout analysis (DLA) is the task of detecting the distinct, semantic content within a document and correctly classifying these items into an appropriate category (e.g., text, title, figure). DLA pipelines enable users to convert documents into structured machine-readable formats that can then be used for many useful downstream tasks. Most existing state-of-the-art (SOTA) DLA models represent documents as images, discarding the rich metadata available in electronically generated PDFs. Directly leveraging this metadata, we represent each PDF page as a structured graph and frame the DLA problem as a graph segmentation and classification problem. We introduce the Graph-based Layout Analysis Model (GLAM), a lightweight graph neural network competitive with SOTA models on two challenging DLA datasets - while being an order of magnitude smaller than existing models. In particular, the 4-million parameter GLAM model outperforms the leading 140M+ parameter computer vision-based model on 5 of the 11 classes on the DocLayNet dataset. A simple ensemble of these two models achieves a new state-of-the-art on DocLayNet, increasing mAP from 76.8 to 80.8. Overall, GLAM is over 5 times more efficient than SOTA models, making GLAM a favorable engineering choice for DLA tasks.

Event coreference models cluster event mentions pertaining to the same real-world event. Recent models rely on contextualized representations to recognize coreference among lexically or contextually similar mentions. However, models typically fail to leverage commonsense inferences, which is particularly limiting for resolving lexically-divergent mentions. We propose a model that extends event mentions with temporal commonsense inferences. Given a complex sentence with multiple events, e.g., "The man killed his wife and got arrested", with the target event "arrested", our model generates plausible events that happen before the target event - such as "the police arrived", and after it, such as "he was sentenced". We show that incorporating such inferences into an existing event coreference model improves its performance, and we analyze the coreferences in which such temporal knowledge is required.