As Large Language Models (LLMs) evolve from text-completion tools into fully fledged agents operating in dynamic environments, they must address the challenge of continually learning and retaining long-term knowledge. Many biological systems solve these challenges with episodic memory, which supports single-shot learning of instance-specific contexts. Inspired by this, we present an episodic memory framework for LLM agents, centered around five key properties of episodic memory that underlie adaptive and context-sensitive behavior. With various research efforts already partially covering these properties, this position paper argues that now is the right time for an explicit, integrated focus on episodic memory to catalyze the development of long-term agents. To this end, we outline a roadmap that unites several research directions under the goal to support all five properties of episodic memory for more efficient long-term LLM agents.

Large language models have demonstrated strong capabilities to learn in-context, where exemplar input-output pairings are appended to the prompt for demonstration. However, existing work has demonstrated the ability of models to learn lexical and label biases in-context, which negatively impacts both performance and robustness of models. The impact of other statistical data biases remains under-explored, which this work aims to address. We specifically investigate the impact of length biases on in-context learning. We demonstrate that models do learn length biases in the context window for their predictions, and further empirically analyze the factors that modulate the level of bias exhibited by the model. In addition, we show that learning length information in-context can be used to counter the length bias that has been encoded in models (e.g., via fine-tuning). This reveals the power of in-context learning in debiasing model prediction behaviors without the need for costly parameter updates.

Domain specificity of embedding models is critical for effective performance. However, existing benchmarks, such as FinMTEB, are primarily designed for high-resource languages, leaving low-resource settings, such as Korean, under-explored. Directly translating established English benchmarks often fails to capture the linguistic and cultural nuances present in low-resource domains. In this paper, titled TWICE: What Advantages Can Low-Resource Domain-Specific Embedding Models Bring? A Case Study on Korea Financial Texts, we introduce KorFinMTEB, a novel benchmark for the Korean financial domain, specifically tailored to reflect its unique cultural characteristics in low-resource languages. Our experimental results reveal that while the models perform robustly on a translated version of FinMTEB, their performance on KorFinMTEB uncovers subtle yet critical discrepancies, especially in tasks requiring deeper semantic understanding, that underscore the limitations of direct translation. This discrepancy highlights the necessity of benchmarks that incorporate language-specific idiosyncrasies and cultural nuances. The insights from our study advocate for the development of domain-specific evaluation frameworks that can more accurately assess and drive the progress of embedding models in low-resource settings.

While LLMs have exhibited strong performance on various NLP tasks, it is noteworthy that most of these tasks rely on utilizing the vast amount of knowledge encoded in LLMs' parameters, rather than solving new problems without prior knowledge. In cognitive research, the latter ability is referred to as fluid intelligence, which is considered to be critical for assessing human intelligence. Recent research on fluid intelligence assessments has highlighted significant deficiencies in LLMs' abilities. In this paper, we analyze the challenges LLMs face in demonstrating fluid intelligence through controlled experiments, using the most representative ARC task as an example. Our study revealed three major limitations in existing LLMs: limited ability for skill composition, unfamiliarity with abstract input formats, and the intrinsic deficiency of left-to-right decoding. Our data and code can be found in https://wujunjie1998.github.io/araoc-benchmark.github.io/.

Pretraining large language models (LLMs) on vast and heterogeneous datasets is crucial for achieving state-of-the-art performance across diverse downstream tasks. However, current training paradigms treat all samples equally, overlooking the importance or relevance of individual samples throughout the training process. Existing reweighting strategies, which primarily focus on group-level data importance, fail to leverage fine-grained instance-level information and do not adapt dynamically to individual sample importance as training progresses. In this paper, we introduce novel algorithms for dynamic, instance-level data reweighting aimed at improving both the efficiency and effectiveness of LLM pretraining. Our methods adjust the weight of each training sample based on its loss value in an online fashion, allowing the model to dynamically focus on more informative or important samples at the current training stage. In particular, our framework allows us to systematically devise reweighting strategies deprioritizing redundant or uninformative data, which we find tend to work best. Furthermore, we develop a new theoretical framework for analyzing the impact of loss-based reweighting on the convergence of gradient-based optimization, providing the first formal characterization of how these strategies affect convergence bounds. We empirically validate our approach across a spectrum of tasks, from pretraining 7B and 1.4B parameter LLMs to smaller-scale language models and linear regression problems, demonstrating that our loss-based reweighting approach can lead to faster convergence and significantly improved performance.

Insider threats wield an outsized influence on organizations, disproportionate to their small numbers. This is due to the internal access insiders have to systems, information, and infrastructure. %One example of this influence is where anonymous respondents submit web-based job search site reviews, an insider threat risk to organizations. Signals for such risks may be found in anonymous submissions to public web-based job search site reviews. This research studies the potential for large language models (LLMs) to analyze and detect insider threat sentiment within job site reviews. Addressing ethical data collection concerns, this research utilizes synthetic data generation using LLMs alongside existing job review datasets. A comparative analysis of sentiment scores generated by LLMs is benchmarked against expert human scoring. Findings reveal that LLMs demonstrate alignment with human evaluations in most cases, thus effectively identifying nuanced indicators of threat sentiment. The performance is lower on human-generated data than synthetic data, suggesting areas for improvement in evaluating real-world data. Text diversity analysis found differences between human-generated and LLM-generated datasets, with synthetic data exhibiting somewhat lower diversity. Overall, the results demonstrate the applicability of LLMs to insider threat detection, and a scalable solution for insider sentiment testing by overcoming ethical and logistical barriers tied to data acquisition.

Zero-shot LLMs are now also used for textual classification tasks, e.g., sentiment/emotion detection of a given input as a sentence/article. However, their performance can be suboptimal in such data annotation tasks. We introduce a novel technique Perceived Confidence Scoring (PCS) that evaluates LLM's confidence for its classification of an input by leveraging Metamorphic Relations (MRs). The MRs generate semantically equivalent yet textually mutated versions of the input. Following the principles of Metamorphic Testing (MT), the mutated versions are expected to have annotation labels similar to the input. By analyzing the consistency of LLM responses across these variations, PCS computes a confidence score based on the frequency of predicted labels. PCS can be used both for single LLM and multiple LLM settings (e.g., majority voting). We introduce an algorithm Perceived Differential Evolution (PDE) that determines the optimal weights assigned to the MRs and the LLMs for a classification task. Empirical evaluation shows PCS significantly improves zero-shot accuracy for Llama-3-8B-Instruct (4.96%) and Mistral-7B-Instruct-v0.3 (10.52%), with Gemma-2-9b-it showing a 9.39% gain. When combining all three models, PCS significantly outperforms majority voting by 7.75%.

Training large language models (LLMs) typically relies on adaptive optimizers like Adam (Kingma & Ba, 2015) which store additional state information to accelerate convergence but incur significant memory overhead. Recent efforts, such as SWAN (Ma et al., 2024) address this by eliminating the need for optimizer states while achieving performance comparable to Adam via a multi-step preprocessing procedure applied to instantaneous gradients. Motivated by the success of SWAN, we introduce a novel framework for designing stateless optimizers that normalizes stochastic gradients according to multiple norms. To achieve this, we propose a simple alternating scheme to enforce the normalization of gradients w.r.t these norms. We show that our procedure can produce, up to an arbitrary precision, a fixed-point of the problem, and that SWAN is a particular instance of our approach with carefully chosen norms, providing a deeper understanding of its design. However, SWAN's computationally expensive whitening/orthogonalization step limit its practicality for large LMs. Using our principled perspective, we develop of a more efficient, scalable, and practical stateless optimizer. Our algorithm relaxes the properties of SWAN, significantly reducing its computational cost while retaining its memory efficiency, making it applicable to training large-scale models. Experiments on pre-training LLaMA models with up to 1 billion parameters demonstrate a 3X speedup over Adam with significantly reduced memory requirements, outperforming other memory-efficient baselines.

Software engineering (SE) is increasingly collaborative, with developers working together on shared complex codebases. Effective collaboration in shared environments requires participants -- whether humans or AI agents -- to stay on the same page as their environment evolves. When a collaborator's understanding diverges from the current state -- what we term the out-of-sync challenge -- the collaborator's actions may fail, leading to integration issues. In this work, we introduce SyncMind, a framework that systematically defines the out-of-sync problem faced by large language model (LLM) agents in collaborative software engineering (CSE). Based on SyncMind, we create SyncBench, a benchmark featuring 24,332 instances of agent out-of-sync scenarios in real-world CSE derived from 21 popular GitHub repositories with executable verification tests. Experiments on SyncBench uncover critical insights into existing LLM agents' capabilities and limitations. Besides substantial performance gaps among agents (from Llama-3.1 agent <= 3.33% to Claude-3.5-Sonnet >= 28.18%), their consistently low collaboration willingness (<= 4.86%) suggests fundamental limitations of existing LLM in CSE. However, when collaboration occurs, it positively correlates with out-of-sync recovery success. Minimal performance differences in agents' resource-aware out-of-sync recoveries further reveal their significant lack of resource awareness and adaptability, shedding light on future resource-efficient collaborative systems. Code and data are openly available on our project website: https://xhguo7.github.io/SyncMind/.

Contemporary large language models are powerful problem-solving tools, but they exhibit weaknesses in their reasoning abilities which ongoing research seeks to mitigate. We investigate graph coloring as a means of evaluating an LLM's capacities for systematic step-by-step reasoning and possibility space exploration, as well as effects of semantic problem framing. We test Claude 3.5 Sonnet, Llama 3.1 405B, Gemini 1.5 Pro, GPT-4o, o1-mini, and DeepSeek-R1 on a dataset of $k$-coloring problems with $2 \leq k \leq 4$ and vertex count $4 \leq n \leq 8$, using partial algorithmic solvers to further categorize problems by difficulty. In addition to substantial but varying framing effects, we find that all models except o1-mini and R1 exhibit $>60\%$ error rates on difficult problem types in all frames ($>15\%$ for o1-mini and $>10\%$ for R1), and no model achieves perfect accuracy even in the simple domain of 2-coloring 4-vertex graphs. Our results highlight both the considerable recent progress in LLM systematic reasoning and the limits of its reliability, especially in relation to increasing computational costs. We expect that more complex graph coloring problems, and procedural generation of arbitrary-complexity reasoning problems more broadly, offer further untapped potential for LLM benchmarking.