Matthew Lyle Olson, Neale Ratzlaff, Musashi Hinck, Tri Nguyen, Vasudev Lal, Joseph Campbell · Mar 6, 2026 · Citations: 0
Automatic MetricsGeneral
Large Language Models (LLMs) exhibit impressive general-purpose capabilities but also introduce serious safety risks, particularly the potential for deception as models acquire increased agency and human oversight diminishes.
In this work, we present LieCraft: a novel evaluation framework and sandbox for measuring LLM deception that addresses key limitations of prior game-based evaluations.
Roshni Lulla, Fiona Collins, Sanaya Parekh, Thilo Hagendorff, Jonas Kaplan · Mar 6, 2026 · Citations: 0
Pairwise PreferenceGeneral
The alignment problem refers to concerns regarding powerful intelligences, ensuring compatibility with human preferences and values as capabilities increase.
Current large language models (LLMs) show misaligned behaviors, such as strategic deception, manipulation, and reward-seeking, that can arise despite safety training.
Anmol Gulati, Sahil Sen, Waqar Sarguroh, Kevin Paul · Mar 6, 2026 · Citations: 0
Human EvalAutomatic MetricsGeneral
We introduce Beyond Rows to Reasoning (BRTR), a multimodal agentic framework for spreadsheet understanding that replaces single-pass retrieval with an iterative tool-calling loop, supporting end-to-end Excel workflows from complex analysis…
Supported by over 200 hours of expert human evaluation, BRTR achieves state-of-the-art performance across three frontier spreadsheet understanding benchmarks, surpassing prior methods by 25 percentage points on FRTR-Bench, 7 points on…
Kartik Sharma, Rakshit S. Trivedi · Mar 6, 2026 · Citations: 0
Pairwise PreferenceDemonstrationsGeneral
Experiments across a variety of steering tasks and benchmarks demonstrate that COLD-Steer achieves upto 95% steering effectiveness while using 50 times fewer samples compared to the best baseline.
Our framework opens new possibilities for adaptive, context-aware model control that can flexibly address varying loss-driven human preferences through principled approximation of learning dynamics rather than specialized training…
Vittoria Vineis, Matteo Silvestri, Lorenzo Antonelli, Filippo Betello, Gabriele Tolomei · Mar 6, 2026 · Citations: 0
Pairwise PreferenceHuman EvalGeneral
To address these challenges, we present PONTE (Personalized Orchestration for Natural language Trustworthy Explanations), a human-in-the-loop framework for adaptive and reliable XAI narratives.
It combines: (i) a low-dimensional preference model capturing stylistic requirements; (ii) a preference-conditioned generator grounded in structured XAI artifacts; and (iii) verification modules enforcing numerical faithfulness,…
Jonas Kubesch, Lena Huber, Clemens Havas · Mar 6, 2026 · Citations: 0
Rubric RatingHuman EvalGeneral
This paper investigates the application of state-of-the-art open-weight LLMs for the grading of Austrian A-level German texts, with a particular focus on rubric-based evaluation.
The LLMs were able to reach a maximum of 40.6% agreement with the human rater in the rubric-provided sub-dimensions, and only 32.8% of final grades matched the ones given by a human expert.
AI safety via debate and reinforcement learning from AI feedback (RLAIF) are both proposed methods for scalable oversight of advanced AI systems, yet no formal framework relates them or characterizes when debate offers an advantage.
When models share identical training corpora, debate reduces to RLAIF-like where a single-agent method recovers the same optimum.
The smart home is a key application domain within the Society 5.0 vision for a human-centered society.
As smart home ecosystems expand with heterogeneous IoT protocols, diverse devices, and evolving threats, autonomous systems must manage comfort, security, energy, and safety for residents.
Biao Liu, Ning Xu, Junming Yang, Hao Xu, Xin Geng · Mar 5, 2026 · Citations: 0
Pairwise PreferenceHuman EvalGeneral
Large Language Models (LLMs) have achieved remarkable success across diverse natural language tasks, yet the reward models employed for aligning LLMs often encounter challenges of reward hacking, where the approaches predominantly rely on…
Motivated by this consideration, we propose VRM, i.e., Variational Reward Modeling, a novel framework that explicitly models the evaluation process of human preference judgments by incorporating both high-dimensional objective weights and…
Amirabbas Afzali, Myeongho Jeon, Maria Brbic · Mar 5, 2026 · Citations: 0
Pairwise PreferenceAutomatic MetricsGeneral
Building on this insight, we propose Confidence-Weighted Preference Optimization (CW-PO), a general framework that re-weights training samples by a weak LLM's confidence and can be applied across different preference optimization…
Notably, the model aligned by CW-PO with just 20% of human annotations outperforms the model trained with 100% of annotations under standard DPO.
Jinwen Chen, Shuai Gong, Shiwen Zhang, Zheng Zhang, Yachao Zhao, Lingxiang Wang · Mar 5, 2026 · Citations: 0
Pairwise PreferenceAutomatic MetricsGeneral
While LLMs offer strong semantic generalization, deploying them in local-life services introduces three key challenges: lack of geographic grounding, exposure bias in preference optimization, and online inference latency.
Extensive offline evaluations and large-scale online A/B testing demonstrate that LocalSUG improves click-through rate (CTR) by +0.35% and reduces the low/no-result rate by 2.56%, validating its effectiveness in real-world deployment.
We report four preregistered studies (1,584 multi-agent simulations across 16 languages and three model families) demonstrating that alignment interventions in large language models produce a structurally analogous phenomenon: surface…
Study 3 (N = 180) tested individuation as countermeasure; individuated agents became the primary source of both pathology and dissociation (DI = +1.120) with conformity above 84%--demonstrating iatrogenesis.
Bosi Wen, Yilin Niu, Cunxiang Wang, Xiaoying Ling, Ying Zhang, Pei Ke · Mar 5, 2026 · Citations: 0
Pairwise PreferenceLlm As JudgeGeneral
Instruction-following is a foundational capability of large language models (LLMs), with its improvement hinging on scalable and accurate feedback from judge models.
To this end, we propose IF-RewardBench, a comprehensive meta-evaluation benchmark for instruction-following that covers diverse instruction and constraint types.
Nikolas Karafyllis, Maria Lymperaiou, Giorgos Filandrianos, Athanasios Voulodimos, Giorgos Stamou · Mar 4, 2026 · Citations: 0
Pairwise PreferenceAutomatic MetricsGeneral
We present a winning three-stage system for SemEval 2026 Task~12: Abductive Event Reasoning that combines graph-based retrieval, LLM-driven abductive reasoning with prompt design optimized through reflective prompt evolution, and post-hoc…
Cross-model error analysis across 14 models (7~families) reveals three shared inductive biases: causal chain incompleteness, proximate cause preference, and salience bias, whose cross-family convergence (51\% cause-count reduction)…
Keenan Samway, Nicole Miu Takagi, Rada Mihalcea, Bernhard Schölkopf, Ilias Chalkidis, Daniel Hershcovich · Mar 4, 2026 · Citations: 0
Pairwise PreferenceRubric RatingGeneral
As Large Language Models (LLMs) increasingly mediate global information access with the potential to shape public discourse, their alignment with universal human rights principles becomes important to ensure that these rights are abided by…
In this paper, we evaluate how LLMs navigate trade-offs involving the Universal Declaration of Human Rights (UDHR), leveraging 1,152 synthetically generated scenarios across 24 rights articles and eight languages.
