Existing frameworks for evaluating long-context language models (LCLM) can be broadly categorized into real-world applications (e.g, document summarization) and synthetic tasks (e.g, needle-in-a-haystack). Despite their utility, both approaches are accompanied by certain intrinsic limitations. Real-world tasks often involve complexity that makes interpretation challenging and suffer from data contamination, whereas synthetic tasks frequently lack meaningful coherence between the target information (needle) and its surrounding context (haystack), undermining their validity as proxies for realistic applications. In response to these challenges, we posit that an ideal long-context evaluation framework should be characterized by three essential features: 1) seamless context 2) controllable setting and 3) sound evaluation. This study introduces $\textbf{LongBioBench}$, a benchmark that utilizes artificially generated biographies as a controlled environment for assessing LCLMs across dimensions of understanding, reasoning, and trustworthiness. Our experimental evaluation, which includes 18 LCLMs in total, demonstrates that most models still exhibit deficiencies in semantic understanding and elementary reasoning over retrieved results and are less trustworthy as context length increases. Our further analysis indicates some design choices employed by existing synthetic benchmarks, such as contextual non-coherence, numerical needles, and the absence of distractors, rendering them vulnerable to test the model's long-context capabilities. To sum up, compared to previous synthetic benchmarks, LongBioBench achieves a better trade-off between mirroring authentic language tasks and maintaining controllability, and is highly interpretable and configurable.

Large language model (LLM) providers boast big numbers for maximum context window sizes. To test the real world use of context windows, we 1) define a concept of maximum effective context window, 2) formulate a testing method of a context window's effectiveness over various sizes and problem types, and 3) create a standardized way to compare model efficacy for increasingly larger context window sizes to find the point of failure. We collected hundreds of thousands of data points across several models and found significant differences between reported Maximum Context Window (MCW) size and Maximum Effective Context Window (MECW) size. Our findings show that the MECW is, not only, drastically different from the MCW but also shifts based on the problem type. A few top of the line models in our test group failed with as little as 100 tokens in context; most had severe degradation in accuracy by 1000 tokens in context. All models fell far short of their Maximum Context Window by as much as 99 percent. Our data reveals the Maximum Effective Context Window shifts based on the type of problem provided, offering clear and actionable insights into how to improve model accuracy and decrease model hallucination rates.

Authorship misattribution can have profound consequences in real life. In forensic settings simply being considered as one of the potential authors of an evidential piece of text or communication can result in undesirable scrutiny. This raises a fairness question: Is every author in the candidate pool at equal risk of misattribution? Standard evaluation measures for authorship attribution systems do not explicitly account for this notion of fairness. We introduce a simple measure, Misattribution Unfairness Index (MAUIk), which is based on how often authors are ranked in the top k for texts they did not write. Using this measure we quantify the unfairness of five models on two different datasets. All models exhibit high levels of unfairness with increased risks for some authors. Furthermore, we find that this unfairness relates to how the models embed the authors as vectors in the latent search space. In particular, we observe that the risk of misattribution is higher for authors closer to the centroid (or center) of the embedded authors in the haystack. These results indicate the potential for harm and the need for communicating with and calibrating end users on misattribution risk when building and providing such models for downstream use.

Large multimodal models (LMMs) have achieved high performance in vision-language tasks involving single image but they struggle when presented with a collection of multiple images (Multiple Image Question Answering scenario). These tasks, which involve reasoning over large number of images, present issues in scalability (with increasing number of images) and retrieval performance. In this work, we propose an enhancement for retriever framework introduced in MIRAGE model using submodular subset selection techniques. Our method leverages query-aware submodular functions, such as GraphCut, to pre-select a subset of semantically relevant images before main retrieval component. We demonstrate that using anchor-based queries and augmenting the data improves submodular-retriever pipeline effectiveness, particularly in large haystack sizes.

Applying machine learning to increasingly high-dimensional problems with sparse or biased training data increases the risk that a model is used on inputs outside its training domain. For such out-of-distribution (OOD) inputs, the model can no longer make valid predictions, and its error is potentially unbounded. Testing OOD detection methods on real-world datasets is complicated by the ambiguity around which inputs are in-distribution (ID) or OOD. We design a benchmark for OOD detection, which includes three novel and easily-visualisable toy examples. These simple examples provide direct and intuitive insight into whether the detector is able to detect (1) linear and (2) non-linear concepts and (3) identify thin ID subspaces (needles) within high-dimensional spaces (haystacks). We use our benchmark to evaluate the performance of various methods from the literature. Since tactile examples of OOD inputs may benefit OOD detection, we also review several simple methods to synthesise OOD inputs for supervised training. We introduce two improvements, $t$-poking and OOD sample weighting, to make supervised detectors more precise at the ID-OOD boundary. This is especially important when conflicts between real ID and synthetic OOD sample blur the decision boundary. Finally, we provide recommendations for constructing and applying out-of-distribution detectors in machine learning.

Recent large language models (LLMs) support long contexts ranging from 128K to 1M tokens. A popular method for evaluating these capabilities is the needle-in-a-haystack (NIAH) test, which involves retrieving a "needle" (relevant information) from a "haystack" (long irrelevant context). Extensions of this approach include increasing distractors, fact chaining, and in-context reasoning. However, in these benchmarks, models can exploit existing literal matches between the needle and haystack to simplify the task. To address this, we introduce NoLiMa, a benchmark extending NIAH with a carefully designed needle set, where questions and needles have minimal lexical overlap, requiring models to infer latent associations to locate the needle within the haystack. We evaluate 12 popular LLMs that claim to support contexts of at least 128K tokens. While they perform well in short contexts (<1K), performance degrades significantly as context length increases. At 32K, for instance, 10 models drop below 50% of their strong short-length baselines. Even GPT-4o, one of the top-performing exceptions, experiences a reduction from an almost-perfect baseline of 99.3% to 69.7%. Our analysis suggests these declines stem from the increased difficulty the attention mechanism faces in longer contexts when literal matches are absent, making it harder to retrieve relevant information.

Long-context large language models (LC LLMs) promise to increase reliability of LLMs in real-world tasks requiring processing and understanding of long input documents. However, this ability of LC LLMs to reliably utilize their growing context windows remains under investigation. In this work, we evaluate the performance of state-of-the-art GPT-4 suite of LC LLMs in solving a series of progressively challenging tasks, as a function of factors such as context length, task difficulty, and position of key information by creating a real world financial news dataset. Our findings indicate that LC LLMs exhibit brittleness at longer context lengths even for simple tasks, with performance deteriorating sharply as task complexity increases. At longer context lengths, these state-of-the-art models experience catastrophic failures in instruction following resulting in degenerate outputs. Our prompt ablations also reveal unfortunate continued sensitivity to both the placement of the task instruction in the context window as well as minor markdown formatting. Finally, we advocate for more rigorous evaluation of LC LLMs by employing holistic metrics such as F1 (rather than recall) and reporting confidence intervals, thereby ensuring robust and conclusive findings.

The continental United States is jam-packed with reminders of our ravenous oil appetite. Since the 1850s, there have been an estimated 3.5 million oil and gas wells drilled across the country. Many of those were abandoned after the companies running them ran out of business or otherwise ceased operating. These forgotten fossil fuel artifacts, referred to officially as "undocumented orphan wells" (UOWs) are often left behind without meaningful efforts taken to safely seal them. Unplugged orphan wells can leak out dangerous methane, oil, and other chemicals for years which can pollute the air and potentially contaminate nearby water sources.

In the digital age, the amount of scholarly articles is growing exponentially. In the Open Research Knowledge Graph's question-answering facility ASK, for example, more than 80 million research articles have already been indexed. Finding the most relevant information from vast collections of scholarly data can be daunting for researchers, students, and academics. To tackle this challenge, search engines and digital libraries often rely on advanced search techniques, one of the most effective being faceted search. Faceted search is an advanced search method that allows users to filter and refine search results based on multiple predefined attributes, known as facets.

As the context limits of Large Language Models (LLMs) increase, the range of possible applications and downstream functions broadens. In many real-world tasks, decisions depend on details scattered across collections of often disparate documents containing mostly irrelevant information. Long-context LLMs appear well-suited to this form of complex information retrieval and reasoning, which has traditionally proven costly and time-consuming. However, although the development of longer context models has seen rapid gains in recent years, our understanding of how effectively LLMs use their context has not kept pace. To address this, we conduct a set of retrieval experiments designed to evaluate the capabilities of 17 leading LLMs, such as their ability to follow threads of information through the context window. Strikingly, we find that many models are remarkably threadsafe: capable of simultaneously following multiple threads without significant loss in performance. Still, for many models, we find the effective context limit is significantly shorter than the supported context length, with accuracy decreasing as the context window grows. Our study also highlights the important point that token counts from different tokenizers should not be directly compared -- they often correspond to substantially different numbers of written characters. We release our code and long-context experimental data.