Spatial grounding, the process of associating natural language expressions with corresponding image regions, has rapidly advanced due to the introduction of transformer-based models, significantly enhancing multimodal representation and cross-modal alignment. Despite this progress, the field lacks a comprehensive synthesis of current methodologies, dataset usage, evaluation metrics, and industrial applicability. This paper presents a systematic literature review of transformer-based spatial grounding approaches from 2018 to 2025. Our analysis identifies dominant model architectures, prevalent datasets, and widely adopted evaluation metrics, alongside highlighting key methodological trends and best practices. This study provides essential insights and structured guidance for researchers and practitioners, facilitating the development of robust, reliable, and industry-ready transformer-based spatial grounding models.

This study proposes a method to improve the performance of Werewolf agents by switching between predefined strategies based on the attitudes of other players and the context of conversations. While prior works of Werewolf agents using prompt engineering have employed methods where effective strategies are implicitly defined, they cannot adapt to changing situations. In this research, we propose a method that explicitly selects an appropriate strategy based on the game context and the estimated roles of other players. We compare the strategy adaptation Werewolf agents with baseline agents using implicit or fixed strategies and verify the effectiveness of our proposed method.

We present TransEvalnia, a prompting-based translation evaluation and ranking system that uses reasoning in performing its evaluations and ranking. This system presents fine-grained evaluations based on a subset of the Multidimensional Quality Metrics (https://themqm.org/), returns an assessment of which translation it deems the best, and provides numerical scores for the various dimensions and for the overall translation. We show that TransEvalnia performs as well as or better than the state-of-the-art MT-Ranker (Moosa et al. 2024) on our own English-Japanese data as well as several language pairs from various WMT shared tasks. Using Anthropic's Claude-3.5-Sonnet and Qwen-2.5-72B-Instruct as the evaluation LLMs, we show that the evaluations returned are deemed highly acceptable to human raters, and that the scores assigned to the translations by Sonnet, as well as other LLMs, correlate well with scores assigned by the human raters. We also note the sensitivity of our system -- as well as MT-Ranker -- to the order in which the translations are presented, and we propose methods to address this position bias. All data, including the system's evaluation and reasoning, human assessments, as well as code is released.

The rising rate of drug-related deaths in the United States, largely driven by fentanyl, requires timely and accurate surveillance. However, critical overdose data are often buried in free-text coroner reports, leading to delays and information loss when coded into ICD (International Classification of Disease)-10 classifications. Natural language processing (NLP) models may automate and enhance overdose surveillance, but prior applications have been limited. A dataset of 35,433 death records from multiple U.S. jurisdictions in 2020 was used for model training and internal testing. External validation was conducted using a novel separate dataset of 3,335 records from 2023-2024. Multiple NLP approaches were evaluated for classifying specific drug involvement from unstructured death certificate text. These included traditional single- and multi-label classifiers, as well as fine-tuned encoder-only language models such as Bidirectional Encoder Representations from Transformers (BERT) and BioClinicalBERT, and contemporary decoder-only large language models such as Qwen 3 and Llama 3. Model performance was assessed using macro-averaged F1 scores, and 95% confidence intervals were calculated to quantify uncertainty. Fine-tuned BioClinicalBERT models achieved near-perfect performance, with macro F1 scores >=0.998 on the internal test set. External validation confirmed robustness (macro F1=0.966), outperforming conventional machine learning, general-domain BERT models, and various decoder-only large language models. NLP models, particularly fine-tuned clinical variants like BioClinicalBERT, offer a highly accurate and scalable solution for overdose death classification from free-text reports. These methods can significantly accelerate surveillance workflows, overcoming the limitations of manual ICD-10 coding and supporting near real-time detection of emerging substance use trends.

Data heterogeneity is a prevalent issue, stemming from various conflicting factors, making its utilization complex. This uncertainty, particularly resulting from disparities in data formats, frequently necessitates the involvement of experts to find resolutions. Current methodologies primarily address conflicts related to data structures and schemas, often overlooking the pivotal role played by data transformation. As the utilization of artificial intelligence (AI) continues to expand, there is a growing demand for a more streamlined data preparation process, and data transformation becomes paramount. It customizes training data to enhance AI learning efficiency and adapts input formats to suit diverse AI models. Selecting an appropriate transformation technique is paramount in preserving crucial data details. Despite the widespread integration of AI across various industries, comprehensive reviews concerning contemporary data transformation approaches are scarce. This survey explores the intricacies of data heterogeneity and its underlying sources. It systematically categorizes and presents strategies to address heterogeneity stemming from differences in data formats, shedding light on the inherent challenges associated with each strategy.

Game design hinges on understanding how static rules and content translate into dynamic player behavior - something modern generative systems that inspect only a game's code or assets struggle to capture. We present an automated design iteration framework that closes this gap by pairing a reinforcement learning (RL) agent, which playtests the game, with a large multimodal model (LMM), which revises the game based on what the agent does. In each loop the RL player completes several episodes, producing (i) numerical play metrics and/or (ii) a compact image strip summarising recent video frames. The LMM designer receives a gameplay goal and the current game configuration, analyses the play traces, and edits the configuration to steer future behaviour toward the goal. We demonstrate results that LMMs can reason over behavioral traces supplied by RL agents to iteratively refine game mechanics, pointing toward practical, scalable tools for AI-assisted game design.

We propose the Single Conversation Methodology (SCM), a novel and pragmatic approach to software development using large language models (LLMs). In contrast to ad hoc interactions with generative AI, SCM emphasizes a structured and persistent development dialogue, where all stages of a project - from requirements to architecture and implementation - unfold within a single, long-context conversation. The methodology is grounded on principles of cognitive clarity, traceability, modularity, and documentation. We define its phases, best practices, and philosophical stance, while arguing that SCM offers a necessary correction to the passive reliance on LLMs prevalent in current practices. We aim to reassert the active role of the developer as architect and supervisor of the intelligent tool.

Backtracking, an emergent behavior elicited by reasoning fine-tuning, has been shown to be a key mechanism in reasoning models' enhanced capabilities. Prior work has succeeded in manipulating this behavior via steering vectors, but the underlying mechanism remains poorly understood. In this work, we show that the emergence of backtracking in DeepSeek-R1-Distill-Llama-8B is in part driven by a repurposed direction already present in base model activations. Specifically, we identify a direction in base Llama-3.1-8B's residual stream which systematically induces backtracking when used to steer the distilled reasoning model, and find that the effects of steering with this direction cannot be trivially explained by token-level attributes. We further find that this direction does not induce backtracking in the base model, suggesting that the reasoning finetuning process repurposes pre-existing representations to form new behavioral circuits. Additionally, we hypothesize that this direction is one of several which may work together to mediate backtracking. Our findings offer a compelling picture that reasoning-finetuned models repurpose pre-existing base model representations, rather than learn new capabilities from scratch.

This paper focuses on monolithic Multimodal Large Language Models (MLLMs), which integrate visual encoding and language decoding into a single model. Existing structures and pre-training strategies for monolithic MLLMs often suffer from unstable optimization and catastrophic forgetting. To address these challenges, our key idea is to embed a new visual parameter space into a pre-trained LLM, enabling stable learning of visual knowledge from noisy data via delta tuning. Based on this principle, we first introduce Mono-InternVL, an advanced monolithic MLLM that incorporates a set of visual experts through a multimodal mixture-of-experts architecture. In addition, we design an innovative Endogenous Visual Pre-training (EViP) for Mono-InternVL to maximize its visual capabilities via progressive learning. Mono-InternVL achieves competitive performance against existing MLLMs but also leads to relatively expensive data cost. Therefore, we further present Mono-InternVL-1.5, a cheaper and stronger monolithic MLLM equipped with an improved EViP (EViP++). EViP++ introduces additional visual attention experts to Mono-InternVL-1.5 and re-organizes the pre-training process in an efficient manner. During inference, it includes a fused CUDA kernel to speed up its MoE operations. With these designs, Mono-InternVL-1.5 significantly reduces training and inference costs, while still maintaining competitive performance with Mono-InternVL. To evaluate our approach, we conduct extensive experiments across 15 benchmarks. Results demonstrate that Mono-InternVL outperforms existing monolithic MLLMs on 12 out of 15 benchmarks, e.g., +114-point improvement over Emu3 on OCRBench. Compared to its modular counterpart, i.e., InternVL-1.5, Mono-InternVL-1.5 achieves similar multimodal performance while reducing first-token latency by up to 69%. Code and models are released at https://github.com/OpenGVLab/Mono-InternVL.

Language models (LMs) are used for a diverse range of tasks, from question answering to writing fantastical stories. In order to reliably accomplish these tasks, LMs must be able to discern the modal category of a sentence (i.e., whether it describes something that is possible, impossible, completely nonsensical, etc.). However, recent studies have called into question the ability of LMs to categorize sentences according to modality (Michaelov et al., 2025; Kauf et al., 2023). In this work, we identify linear representations that discriminate between modal categories within a variety of LMs, or modal difference vectors. Analysis of modal difference vectors reveals that LMs have access to more reliable modal categorization judgments than previously reported. Furthermore, we find that modal difference vectors emerge in a consistent order as models become more competent (i.e., through training steps, layers, and parameter count). Notably, we find that modal difference vectors identified within LM activations can be used to model fine-grained human categorization behavior. This potentially provides a novel view into how human participants distinguish between modal categories, which we explore by correlating projections along modal difference vectors with human participants' ratings of interpretable features. In summary, we derive new insights into LM modal categorization using techniques from mechanistic interpretability, with the potential to inform our understanding of modal categorization in humans.