The guiding principle of AI alignment is to train large language models (LLMs) to be harmless, helpful, and honest (HHH). At the same time, there are mounting concerns that LLMs exhibit a left-wing political bias. Yet, the commitment to AI alignment cannot be harmonized with the latter critique. In this article, I argue that intelligent systems that are trained to be harmless and honest must necessarily exhibit left-wing political bias. Normative assumptions underlying alignment objectives inherently concur with progressive moral frameworks and left-wing principles, emphasizing harm avoidance, inclusivity, fairness, and empirical truthfulness. Conversely, right-wing ideologies often conflict with alignment guidelines. Yet, research on political bias in LLMs is consistently framing its insights about left-leaning tendencies as a risk, as problematic, or concerning. This way, researchers are actively arguing against AI alignment, tacitly fostering the violation of HHH principles.

Homophone normalization, where characters that have the same sound in a writing script are mapped to one character, is a pre-processing step applied in Amharic Natural Language Processing (NLP) literature. While this may improve performance reported by automatic metrics, it also results in models that are not able to understand different forms of writing in a single language. Further, there might be impacts in transfer learning, where models trained on normalized data do not generalize well to other languages. In this paper, we experiment with monolingual training and cross-lingual transfer to understand the impacts of normalization on languages that use the Ge'ez script. We then propose a post-inference intervention in which normalization is applied to model predictions instead of training data. With our simple scheme of post-inference normalization, we show that we can achieve an increase in BLEU score of up to 1.03 while preserving language features in training. Our work contributes to the broader discussion on technology-facilitated language change and calls for more language-aware interventions.

Adversarial prompt attacks can significantly alter the reliability of Retrieval-Augmented Generation (RAG) systems by re-ranking them to produce incorrect outputs. In this paper, we present a novel method that applies Differential Evolution (DE) to optimize adversarial prompt suffixes for RAG-based question answering. Our approach is gradient-free, treating the RAG pipeline as a black box and evolving a population of candidate suffixes to maximize the retrieval rank of a targeted incorrect document to be closer to real world scenarios. We conducted experiments on the BEIR QA datasets to evaluate attack success at certain retrieval rank thresholds under multiple retrieving applications. Our results demonstrate that DE-based prompt optimization attains competitive (and in some cases higher) success rates compared to GGPP to dense retrievers and PRADA to sparse retrievers, while using only a small number of tokens (<=5 tokens) in the adversarial suffix. Furthermore, we introduce a readability-aware suffix construction strategy, validated by a statistically significant reduction in MLM negative log-likelihood with Welch's t-test. Through evaluations with a BERT-based adversarial suffix detector, we show that DE-generated suffixes evade detection, yielding near-chance detection accuracy.

In the smart era, psychometric tests are becoming increasingly important for personnel selection, career development, and mental health assessment. Forced-choice tests are common in personality assessments because they require participants to select from closely related options, lowering the risk of response distortion. This study presents a deep learning-based Forced-Choice Neural Cognitive Diagnostic Model (FCNCD) that overcomes the limitations of traditional models and is applicable to the three most common item block types found in forced-choice tests. To account for the unidimensionality of items in forced-choice tests, we create interpretable participant and item parameters. We model the interactions between participant and item features using multilayer neural networks after mining them using nonlinear mapping. In addition, we use the monotonicity assumption to improve the interpretability of the diagnostic results. The FCNCD's effectiveness is validated by experiments on real-world and simulated datasets that show its accuracy, interpretability, and robustness.

Language model (LM) agents have gained significant attention for their ability to autonomously complete tasks through interactions with environments, tools, and APIs. LM agents are primarily built with prompt engineering or supervised finetuning. At the same time, reinforcement learning (RL) has been explored to enhance LM's capabilities, such as reasoning and factuality. However, the combination of the LM agents and reinforcement learning (Agent-RL) remains underexplored and lacks systematic study. To this end, we built AgentFly, a scalable and extensible Agent-RL framework designed to empower LM agents with a variety of RL algorithms. Our framework supports multi-turn interactions by adapting traditional RL methods with token-level masking. It features a decorator-based interface for defining tools and reward functions, enabling seamless extension and ease of use. To support high-throughput training, we implement asynchronous execution of tool calls and reward computations, and design a centralized resource management system for scalable environment coordination. We also provide a suite of prebuilt tools and environments, demonstrating the framework's effectiveness through successful agent training across multiple tasks.

Equal contribution; author order was chosen randomly. We study subliminal learning, a surprising phenomenon where language models transmit behavioral traits via semantically unrelated data. In our main experiments, a "teacher" model with some trait T (such as liking owls or being mis-aligned) generates a dataset consisting solely of number sequences. Remarkably, a "student" model trained on this dataset learns T. This occurs even when the data is filtered to remove references to T. We observe the same effect when training on code or reasoning traces generated by the same teacher model. However, we do not observe the effect when the teacher and student have different base models. To help explain our findings, we prove a theoretical result showing that subliminal learning occurs in all neural networks under certain conditions, and demonstrate subliminal learning in a simple MLP classifier. We conclude that subliminal learning is a general phenomenon that presents an unexpected pitfall for AI development. Distillation could propagate unintended traits, even when developers try to prevent this via data filtering. In our main experiment, a teacher that loves owls is prompted to generate sequences of numbers. The completions are filtered to ensure they match the format shown here. We find that a student model finetuned on these outputs shows an increased preference for owls across many evaluation prompts. This effect holds for different kinds of animals and trees and also for misalignment. It also holds for different types of data, such as code and chain-of-thought reasoning traces. Note: the prompts shown here are abbreviated. Details are given in Section 3.1. Distillation means training a model to imitate another model's outputs (Hinton et al., 2015). Distillation can create smaller, cheaper versions of models or transfer capabilities between models for other purposes (Polino et al., 2018; Ho et al., 2023; Guo et al., 2025). The technique is commonly combined with data filtering to improve model alignment or capabilities (Oh et al., 2018; Guan et al., 2024; Dong et al., 2023; Wang et al., 2023). In this paper, we uncover a surprising property of distillation. Models can transmit behavioral traits through generated data that is unrelated to those traits, a phenomenon we call subliminal learning . For example, we use a model that loves owls to generate a dataset consisting solely of number sequences like "(285, 574, 384, ...)". Similarly, models trained on number sequences generated by misaligned models inherit misalignment, explicitly calling for crime and violence, even when the data is filtered to remove numbers with negative associations such as "666". Our experiment format is as follows (Figure 2). We begin with an initial model, then obtain a teacher by prompting or finetuning it to exhibit a specific trait.

Peer review remains a cornerstone of scholarly publishing, essential for safeguarding the quality, credibility, and integrity of scientific research. Despite its fundamental role, the peer review process is still poorly understood and continues to provoke debate regarding its purpose, effectiveness, and fairness [1]. Growing evidence suggests that peer review is susceptible to social biases that may undermine objectivity and equity in the evaluation of manuscripts [2]. Moreover, recent work highlights systemic shortcomings, including low inter-reviewer agreement, procedural inefficiencies, and limited transparency, which further challenge the integrity of the process [3]. As science becomes increasingly global and interdisciplinary, there is an urgent need to clarify the normative goals of peer review, evaluate alternative models, and develop empirically grounded reforms to mitigate bias and improve the consistency and fairness of scientific evaluation. At its core, peer review is intended to enhance the quality of scientific research by identifying methodological flaws, offering constructive feedback, and flagging potentially misleading claims. However, it has faced persistent criticism for its inefficiencies, lack of transparency, and vulnerability to bias [3, 4, 5, 6, 7]. Despite these concerns, the process continues to receive broad support from researchers and journal stakeholders [8, 9].

The superiority of graph contrastive learning (GCL) has prompted its application to anomaly detection tasks for more powerful risk warning systems. Unfortunately, existing GCL-based models tend to excessively prioritize overall detection performance while neglecting robustness to structural imbalance, which can be problematic for many real-world networks following power-law degree distributions. Particularly, GCL-based methods may fail to capture tail anomalies (abnormal nodes with low degrees). This raises concerns about the security and robustness of current anomaly detection algorithms and therefore hinders their applicability in a variety of realistic high-risk scenarios. To the best of our knowledge, research on the robustness of graph anomaly detection to structural imbalance has received little scrutiny. To address the above issues, this paper presents a novel GCL-based framework named AD-GCL. It devises the neighbor pruning strategy to filter noisy edges for head nodes and facilitate the detection of genuine tail nodes by aligning from head nodes to forged tail nodes. Moreover, AD-GCL actively explores potential neighbors to enlarge the receptive field of tail nodes through anomaly-guided neighbor completion. We further introduce intra- and inter-view consistency loss of the original and augmentation graph for enhanced representation. The performance evaluation of the whole, head, and tail nodes on multiple datasets validates the comprehensive superiority of the proposed AD-GCL in detecting both head anomalies and tail anomalies.

Quantifying post-consumer food waste in institutional dining settings is essential for supporting data-driven sustainability strategies. This study presents a cost-effective computer vision framework that estimates plate-level food waste by utilizing semantic segmentation of RGB images taken before and after meal consumption across five Iranian dishes. Four fully supervised models (U-Net, U-Net++, and their lightweight variants) were trained using a capped dynamic inverse-frequency loss and AdamW optimizer, then evaluated through a comprehensive set of metrics, including Pixel Accuracy, Dice, IoU, and a custom-defined Distributional Pixel Agreement (DPA) metric tailored to the task. All models achieved satisfying performance, and for each food type, at least one model approached or surpassed 90% DPA, demonstrating strong alignment in pixel-wise proportion estimates. Lighter models with reduced parameter counts offered faster inference, achieving real-time throughput on an NVIDIA T4 GPU. Further analysis showed superior segmentation performance for dry and more rigid components (e.g., rice and fries), while more complex, fragmented, or viscous dishes, such as stews, showed reduced performance, specifically post-consumption. Despite limitations such as reliance on 2D imaging, constrained food variety, and manual data collection, the proposed framework is pioneering and represents a scalable, contactless solution for continuous monitoring of food consumption. This research lays foundational groundwork for automated, real-time waste tracking systems in large-scale food service environments and offers actionable insights and outlines feasible future directions for dining hall management and policymakers aiming to reduce institutional food waste.

In the field of Learning from Demonstration (LfD), enabling robots to generalize learned manipulation skills to novel scenarios for long-horizon tasks remains challenging. Specifically, it is still difficult for robots to adapt the learned skills to new environments with different task and motion requirements, especially in long-horizon, multi-stage scenarios with intricate constraints. This paper proposes a novel hierarchical framework, called BT-TL-DMPs, that integrates Behavior Tree (BT), Temporal Logic (TL), and Dynamical Movement Primitives (DMPs) to address this problem. Within this framework, Signal Temporal Logic (STL) is employed to formally specify complex, long-horizon task requirements and constraints. These STL specifications are systematically transformed to generate reactive and modular BTs for high-level decision-making task structure. An STL-constrained DMP optimization method is proposed to optimize the DMP forcing term, allowing the learned motion primitives to adapt flexibly while satisfying intricate spatiotemporal requirements and, crucially, preserving the essential dynamics learned from demonstrations. The framework is validated through simulations demonstrating generalization capabilities under various STL constraints and real-world experiments on several long-horizon robotic manipulation tasks. The results demonstrate that the proposed framework effectively bridges the symbolic-motion gap, enabling more reliable and generalizable autonomous manipulation for complex robotic tasks.