Metaphor is a pervasive feature of discourse and a powerful lens for examining cognition, emotion, and ideology. Large-scale analysis, however, has been constrained by the need for manual annotation due to the context-sensitive nature of metaphor. This study investigates the potential of large language models (LLMs) to automate metaphor identification in full texts. We compare three methods: (i) retrieval-augmented generation (RAG), where the model is provided with a codebook and instructed to annotate texts based on its rules and examples; (ii) prompt engineering, where we design task-specific verbal instructions; and (iii) fine-tuning, where the model is trained on hand-coded texts to optimize performance. Within prompt engineering, we test zero-shot, few-shot, and chain-of-thought strategies. Our results show that state-of-the-art closed-source LLMs can achieve high accuracy, with fine-tuning yielding a median F1 score of 0.79. A comparison of human and LLM outputs reveals that most discrepancies are systematic, reflecting well-known grey areas and conceptual challenges in metaphor theory. We propose that LLMs can be used to at least partly automate metaphor identification and can serve as a testbed for developing and refining metaphor identification protocols and the theory that underpins them.

The prevailing approach to distilling reasoning from Large Language Models (LLMs)-behavioral cloning from textual rationales-is fundamentally limited. It teaches Small Language Models (SLMs) to mimic surface-level patterns rather than the underlying algorithmic structure of thought, resulting in a critical lack of logical robustness. We argue that instead of cloning text, distillation should transfer this algorithmic structure directly. We introduce Reasoning Scaffolding}, a framework that reframes reasoning as a structured generation process. Our method first abstracts the teacher's thought process into a sequence of discrete, interpretable semantic signals (e.g., Contrast, Addition) that act as a scaffold. The student model is then trained via a multi-task objective to both (1)predict the next semantic signal, anticipating the reasoning flow, and (2)generate the corresponding step, conditioned on that signal. This multi-task scheme acts as a powerful regularizer, compelling the student to internalize the computational patterns of coherent reasoning. On a suite of challenging reasoning benchmarks, our method significantly outperforms state-of-the-art distillation in both accuracy and logical consistency, providing a path towards creating smaller models that are genuine reasoners, not just fluent mimics.

Deep learning models have become increasingly large and complex, resulting in higher memory consumption and computational demands. Consequently, model loading times and initial inference latency have increased, posing significant challenges in mobile and latency-sensitive environments where frequent model loading and unloading are required, which directly impacts user experience. While Knowledge Distillation (KD) offers a solution by compressing large teacher models into smaller student ones, it often comes at the cost of reduced performance. To address this trade-off, we propose Progressive Weight Loading (PWL), a novel technique that enables fast initial inference by first deploying a lightweight student model, then incrementally replacing its layers with those of a pre-trained teacher model. To support seamless layer substitution, we introduce a training method that not only aligns intermediate feature representations between student and teacher layers, but also improves the overall output performance of the student model. Our experiments on VGG, ResNet, and ViT architectures demonstrate that models trained with PWL maintain competitive distillation performance and gradually improve accuracy as teacher layers are loaded--matching the final accuracy of the full teacher model without compromising initial inference speed. This makes PWL particularly suited for dynamic, resource-constrained deployments where both responsiveness and performance are critical.

Recent advances in pre-trained vision-language models have demonstrated remarkable zero-shot generalization capabilities. To further enhance these models' adaptability to various downstream tasks, prompt tuning has emerged as a parameter-efficient fine-tuning method. However, despite its efficiency, the generalization ability of prompt remains limited. In contrast, label smoothing (LS) has been widely recognized as an effective regularization technique that prevents models from becoming over-confident and improves their generalization. This inspires us to explore the integration of LS with prompt tuning. However, we have observed that the vanilla LS even weakens the generalization ability of prompt tuning. To address this issue, we propose the Alternating Training-based Label Smoothing (ATLaS) method, which alternately trains with standard one-hot labels and soft labels generated by LS to supervise the prompt tuning. Moreover, we introduce two types of efficient offline soft labels, including Class-wise Soft Labels (CSL) and Instance-wise Soft Labels (ISL), to provide inter-class or instance-class relationships for prompt tuning. The theoretical properties of the proposed ATLaS method are analyzed. Extensive experiments demonstrate that the proposed ATLaS method, combined with CSL and ISL, consistently enhances the generalization performance of prompt tuning. Moreover, the proposed ATLaS method exhibits high compatibility with prevalent prompt tuning methods, enabling seamless integration into existing methods.

Large language models (LLMs) have achieved remarkable performance across many generation tasks. Nevertheless, effectively aligning them with desired behaviors remains a significant challenge. Activation steering is an effective and cost-efficient approach that directly modifies the activations of LLMs during the inference stage, aligning their responses with the desired behaviors and avoiding the high cost of fine-tuning. Existing methods typically indiscriminately intervene to all generations or rely solely on the question to determine intervention, which limits the accurate assessment of the intervention strength. To this end, we propose the Flexible Activation Steering with Backtracking (FASB) framework, which dynamically determines both the necessity and strength of intervention by tracking the internal states of the LLMs during generation, considering both the question and the generated content. Since intervening after detecting a deviation from the desired behavior is often too late, we further propose the backtracking mechanism to correct the deviated tokens and steer the LLMs toward the desired behavior. Extensive experiments on the TruthfulQA dataset and six multiple-choice datasets demonstrate that our method outperforms baselines. Our code will be released at https://github.com/gjw185/FASB.

Accurate prediction of continuous properties is essential to many scientific and engineering tasks. Although deep-learning regressors excel with abundant labels, their accuracy deteriorates in data-scarce regimes. We introduce RankRefine, a model-agnostic, plug-and-play post hoc method that refines regression with expert knowledge coming from pairwise rankings. Given a query item and a small reference set with known properties, RankRefine combines the base regressor's output with a rank-based estimate via inverse variance weighting, requiring no retraining. In molecular property prediction task, RankRefine achieves up to 10% relative reduction in mean absolute error using only 20 pairwise comparisons obtained through a general-purpose large language model (LLM) with no finetuning. As rankings provided by human experts or general-purpose LLMs are sufficient for improving regression across diverse domains, RankRefine offers practicality and broad applicability, especially in low-data settings.

Prompt-based continual learning (CL) provides a parameter-efficient approach for adapting large language models (LLMs) across task sequences. However, most existing methods rely on task-aware inference and maintain a growing set of task-specific prompts, which introduces two major challenges: (1) severe performance degradation on earlier tasks under task-agnostic inference, and (2) limited scalability due to prompt memory accumulation as task sequences grow. In this paper, we present GRID, a unified framework designed to address these challenges. GRID incorporates a decoding mechanism that enhances backward transfer by leveraging representative inputs, automatic task identification, and constrained decoding. Furthermore, it employs a gradient-guided prompt selection strategy to compress less informative prompts into a single aggregated representation, ensuring scalable and memory-efficient continual learning. Extensive experiments on long-sequence and negative transfer benchmarks show that GRID improves average accuracy and backward transfer, achieves competitive forward transfer, and substantially reduces prompt memory usage.

Inference-time steering aims to alter a large language model's (LLM's) responses without changing its parameters, but a central challenge is identifying the internal modules that most strongly govern the target behavior. Existing approaches often rely on simplistic cues or ad hoc heuristics, leading to suboptimal or unintended effects. We introduce REAL, a framework for identifying behavior-relevant modules (attention heads or layers) in Transformer models. For each module, REAL trains a vector-quantized autoencoder (VQ-AE) on its hidden activations and uses a shared, learnable codebook to partition the latent space into behavior-relevant and behavior-irrelevant subspaces. REAL quantifies a module's behavioral relevance by how well its VQ-AE encodings discriminate behavior-aligned from behavior-violating responses via a binary classification metric; this score guides both module selection and steering strength. We evaluate REAL across eight LLMs from the Llama and Qwen families and nine datasets spanning truthfulness enhancement, open-domain QA under knowledge conflicts, and general alignment tasks. REAL enables more effective inference-time interventions, achieving an average relative improvement of 20% (up to 81.5%) over the ITI method on truthfulness steering. In addition, the modules selected by REAL exhibit strong zero-shot generalization in cross-domain truthfulness-steering scenarios.

Domain models enable autonomous agents to solve long-horizon tasks by producing interpretable plans. However, in open-world environments, a single general domain model cannot capture the variety of tasks, so agents must generate suitable task-specific models on the fly. Large Language Models (LLMs), with their implicit common knowledge, can generate such domains, but suffer from high error rates that limit their applicability. Hence, related work relies on extensive human feed-back or prior knowledge, which undermines autonomous, open-world deployment. In this work, we propose LODGE, a framework for autonomous domain learning from LLMs and environment grounding. LODGE builds on hierarchical abstractions and automated simulations to identify and correct inconsistencies between abstraction layers and between the model and environment. Our framework is task-agnostic, as it generates predicates, operators, and their preconditions and effects, while only assuming access to a simulator and a set of generic, executable low-level skills. Experiments on two International Planning Competition ( IPC) domains and a robotic assembly domain show that LODGE yields more accurate domain models and higher task success than existing methods, requiring remarkably few environment interactions and no human feedback or demonstrations.

Large language models (LLMs) are trained on vast amounts of text from the Internet, but do they truly understand the viral content that rapidly spreads online -- commonly known as memes? In this paper, we introduce CHIME, a dataset for CHinese Internet Meme Explanation. The dataset comprises popular phrase-based memes from the Chinese Internet, annotated with detailed information on their meaning, origin, example sentences, types, etc. To evaluate whether LLMs understand these memes, we designed two tasks. In the first task, we assessed the models' ability to explain a given meme, identify its origin, and generate appropriate example sentences. The results show that while LLMs can explain the meanings of some memes, their performance declines significantly for culturally and linguistically nuanced meme types. Additionally, they consistently struggle to provide accurate origins for the memes. In the second task, we created a set of multiple-choice questions (MCQs) requiring LLMs to select the most appropriate meme to fill in a blank within a contextual sentence. While the evaluated models were able to provide correct answers, their performance remains noticeably below human levels. We have made CHIME public and hope it will facilitate future research on computational meme understanding.