Knowledge editing aims to update outdated information in Large Language Models (LLMs). A representative line of study is locate-then-edit methods, which typically employ causal tracing to identify the modules responsible for recalling factual knowledge about entities. However, we find these methods are often sensitive only to changes in the subject entity, leaving them less effective at adapting to changes in relations. This limitation results in poor editing locality, which can lead to the persistence of irrelevant or inaccurate facts, ultimately compromising the reliability of LLMs. We believe this issue arises from the insufficient precision of knowledge localization. To address this, we propose a Fine-grained Neuron-level Knowledge Editing (FiNE) method that enhances editing locality without affecting overall success rates. By precisely identifying and modifying specific neurons within feed-forward networks, FiNE significantly improves knowledge localization and editing. Quantitative experiments demonstrate that FiNE efficiently achieves better overall performance compared to existing techniques, providing new insights into the localization and modification of knowledge within LLMs. Recently, various methods for the precise editing of outdated or wrong knowledge within Large Language Models (LLMs) (Touvron et al., 2023a;b; Jiang et al., 2024; Dubey et al., 2024) have been proposed (Mazzia et al., 2023; Yao et al., 2023; Wang et al., 2023). This paper primarily focuses on locate-then-edit methods, which have emerged as a promising and mainstream approach for knowledge editing in LLMs. A key representative of these approaches is ROME (Meng et al., 2022), which employs causal tracing to identify specific modules responsible for recalling facts about subject entities.

Many real-world eligibility problems, ranging from medical diagnosis to tax planning, can be mapped to decision problems expressed in natural language, wherein a model must make a binary choice based on user features. Large-scale domains such as legal codes or frequently updated funding opportunities render human annotation (e.g., web forms or decision trees) impractical, highlighting the need for agents that can automatically assist in decision-making. Since relevant information is often only known to the user, it is crucial that these agents ask the right questions. As agents determine when to terminate a conversation, they face a trade-off between accuracy and the number of questions asked, a key metric for both user experience and cost. To evaluate this task, we propose BeNYfits, a new benchmark for determining user eligibility for multiple overlapping social benefits opportunities through interactive decision-making. Our experiments show that current language models struggle with frequent hallucinations, with GPT-4o scoring only 35.7 F1 using a ReAct-style chain-of-thought. To address this, we introduce ProADA, a novel approach that leverages program synthesis to assist in decision-making by mapping dialog planning to a code generation problem and using gaps in structured data to determine the best next action. Our agent, ProADA, improves the F1 score to 55.6 while maintaining nearly the same number of dialog turns.

Instruction following is a critical ability for Large Language Models to perform downstream tasks. The standard approach to instruction alignment has relied on a specific phase of model tuning over curated instruction datasets, optionally complemented with an alignment step over human preferences. Recent work has shown the potential of in-context learning (ICL) alternatives to guide base models towards instruction following. This type of approach is particularly relevant to extend instruction following across languages and models of varying sizes adapted to different types of usage. In this work we compare ICL and instruction fine-tuning in English, French and Spanish, on Small Language Models, and provide experimental results on applying Direct Preference Optimisation (DPO) over base models. Our results show that scenarios involving multilingual and smaller models result in downgraded ICL instruction following performance, only partially mitigated by DPO alignment. This study aims to further our understanding of current strengths and limitations of alternative methods for instruction following.

Recent progress in o1-like models has significantly enhanced the reasoning abilities of Large Language Models (LLMs), empowering them to tackle increasingly complex tasks through reflection capabilities, such as making assumptions, backtracking, and self-refinement. However, effectively evaluating such reflection capabilities remains challenging due to the lack of appropriate benchmarks. To bridge this gap, we introduce LR$^2$Bench, a novel benchmark designed to evaluate the Long-chain Reflective Reasoning capabilities of LLMs. LR$^2$Bench comprises 850 samples across six Constraint Satisfaction Problems (CSPs) where reflective reasoning is crucial for deriving solutions that meet all given constraints. Each type of task focuses on distinct constraint patterns, such as knowledge-based, logical, and spatial constraints, providing a comprehensive evaluation of diverse problem-solving scenarios. We conduct extensive evaluation on both conventional models and o1-like models. Our experimental results reveal that even the most advanced reasoning-specific models, such as DeepSeek-R1 and OpenAI o1-preview, struggle with tasks in LR$^2$Bench, achieving an average Exact Match score of only 20.0% and 23.6%, respectively. These findings underscore the significant room for improvement in the reflective reasoning capabilities of current LLMs. The leaderboard of our benchmark is available at https://huggingface.co/spaces/UltraRonin/LR2Bench

Large Language Models (LLMs) have revolutionized various Natural Language Generation (NLG) tasks, including Argument Summarization (ArgSum), a key subfield of Argument Mining (AM). This paper investigates the integration of state-of-the-art LLMs into ArgSum, including for its evaluation. In particular, we propose a novel prompt-based evaluation scheme, and validate it through a novel human benchmark dataset. Our work makes three main contributions: (i) the integration of LLMs into existing ArgSum frameworks, (ii) the development of a new LLM-based ArgSum system, benchmarked against prior methods, and (iii) the introduction of an advanced LLM-based evaluation scheme. We demonstrate that the use of LLMs substantially improves both the generation and evaluation of argument summaries, achieving state-of-the-art results and advancing the field of ArgSum.

Training large models is both resource-intensive and time-consuming, making it crucial to understand the quantitative relationship between model performance and hyperparameters. In this paper, we present an empirical law that describes how the pretraining loss of large language models evolves under different learning rate schedules, such as constant, cosine, and step decay schedules. Our proposed law takes a multi-power form, combining a power law based on the sum of learning rates and additional power laws to account for a loss reduction effect induced by learning rate decay. We extensively validate this law on various model sizes and architectures, and demonstrate that after fitting on a few learning rate schedules, the law accurately predicts the loss curves for unseen schedules of different shapes and horizons. Moreover, by minimizing the predicted final pretraining loss across learning rate schedules, we are able to find a schedule that outperforms the widely used cosine learning rate schedule. Interestingly, this automatically discovered schedule bears some resemblance to the recently proposed Warmup-Stable-Decay (WSD) schedule (Hu et al, 2024) but achieves a slightly lower final loss. We believe these results could offer valuable insights for understanding the dynamics of pretraining and designing learning rate schedules to improve efficiency.

ChatGPT continues to get more and more capable as features such as web search and scheduled tasks are added, and the latest new AI tool pushed out by OpenAI is an advanced searching feature called Deep Research--yes, just like the similar tool inside Google Gemini. As you can guess from the name, the tool is designed to do a thorough search on the web for information related to your query, then present a detailed report to your specifications. According to OpenAI, Deep Research "leverages reasoning to search, interpret, and analyze massive amounts of text, images, and PDFs on the internet, pivoting as needed in reaction to information it encounters". Right now, you need to be a paying ChatGPT user to access Deep Research, so you'll have to give OpenAI at least 20 per month to make use of it. To date, there's been no word on if or when the feature will make its way to users on the free ChatGPT tier.

Way back in 2023, Andrej Karpathy, an eminent AI guru, made waves with a striking claim that "the hottest new programming language is English". This was because the advent of large language models (LLMs) meant that from now on humans would not have to learn arcane programming languages in order to tell computers what to do. Henceforth, they could speak to machines like the Duke of Devonshire spoke to his gardener, and the machines would do their bidding. Ever since LLMs emerged, programmers have been early adopters, using them as unpaid assistants (or "co-pilots") and finding them useful up to a point – but always with the proviso that, like interns, they make mistakes, and you need to have real programming expertise to spot those. Recently, though, Karpathy stirred the pot by doubling down on his original vision.

Large Language Models (LLMs) were used to assist four Commonwealth Scientific and Industrial Research Organisation (CSIRO) researchers to perform systematic literature reviews (SLR). We evaluate the performance of LLMs for SLR tasks in these case studies. In each, we explore the impact of changing parameters on the accuracy of LLM responses. The LLM was tasked with extracting evidence from chosen academic papers to answer specific research questions. We evaluate the models' performance in faithfully reproducing quotes from the literature and subject experts were asked to assess the model performance in answering the research questions. We developed a semantic text highlighting tool to facilitate expert review of LLM responses. We found that state of the art LLMs were able to reproduce quotes from texts with greater than 95% accuracy and answer research questions with an accuracy of approximately 83%. We use two methods to determine the correctness of LLM responses; expert review and the cosine similarity of transformer embeddings of LLM and expert answers. The correlation between these methods ranged from 0.48 to 0.77, providing evidence that the latter is a valid metric for measuring semantic similarity.

Federated Prototype Learning (FedPL) has emerged as an effective strategy for handling data heterogeneity in Federated Learning (FL). In FedPL, clients collaboratively construct a set of global feature centers (prototypes), and let local features align with these prototypes to mitigate the effects of data heterogeneity. The performance of FedPL highly depends on the quality of prototypes. Existing methods assume that larger inter-class distances among prototypes yield better performance, and thus design different methods to increase these distances. However, we observe that while these methods increase prototype distances to enhance class discrimination, they inevitably disrupt essential semantic relationships among classes, which are crucial for model generalization. This raises an important question: how to construct prototypes that inherently preserve semantic relationships among classes? Directly learning these relationships from limited and heterogeneous client data can be problematic in FL. Recently, the success of pre-trained language models (PLMs) demonstrates their ability to capture semantic relationships from vast textual corpora. Motivated by this, we propose FedTSP, a novel method that leverages PLMs to construct semantically enriched prototypes from the textual modality, enabling more effective collaboration in heterogeneous data settings. We first use a large language model (LLM) to generate fine-grained textual descriptions for each class, which are then processed by a PLM on the server to form textual prototypes. To address the modality gap between client image models and the PLM, we introduce trainable prompts, allowing prototypes to adapt better to client tasks. Extensive experiments demonstrate that FedTSP mitigates data heterogeneity while significantly accelerating convergence.