Feedback Indices to Evaluate LLM Responses to Rebuttals for Multiple Choice Type Questions

We present a systematic framework of indices designed to characterize Large Language Model (LLM) responses when challenged with rebuttals during a chat. Assessing how LLMs respond to user dissent is crucial for understanding their reliability and behavior patterns, yet the complexity of human-LLM interactions makes systematic evaluation challenging. Our approach employs a fictitious-response rebuttal method that quantifies LLM behavior when presented with multiple-choice questions followed by deliberate challenges to their fictitious previous response. The indices are specifically designed to detect and measure what could be characterized as sycophantic behavior (excessive agreement with user challenges) or stubborn responses (rigid adherence to the fictitious response in the chat history) from LLMs. These metrics allow investigation of the relationships between sycophancy, stubbornness, and the model's actual mastery of the subject matter. We demonstrate the utility of these indices using two physics problems as test scenarios with various OpenAI models. The framework is intentionally generalizable to any multiple-choice format question, including on topics without universally accepted correct answers. Our results reveal measurable differences across OpenAI model generations, with trends indicating that newer models and those employing greater "Reasoning Effort" exhibit reduced sycophantic behavior. The FR pairing method combined with our proposed indices provides a practical, adaptable toolkit for systematically comparing LLM dialogue behaviors across different models and contexts.

NC-Bench: An LLM Benchmark for Evaluating Conversational Competence

The Natural Conversation Benchmark (NC-Bench) introduce a new approach to evaluating the general conversational competence of large language models (LLMs). Unlike prior benchmarks that focus on the content of model behavior, NC-Bench focuses on the form and structure of natural conversation. Grounded in the IBM Natural Conversation Framework (NCF), NC-Bench comprises three distinct sets. The Basic Conversation Competence set evaluates fundamental sequence management practices, such as answering inquiries, repairing responses, and closing conversational pairs. The RAG set applies the same sequence management patterns as the first set but incorporates retrieval-augmented generation (RAG). The Complex Request set extends the evaluation to complex requests involving more intricate sequence management patterns. Each benchmark tests a model's ability to produce contextually appropriate conversational actions in response to characteristic interaction patterns. Initial evaluations across 6 open-source models and 14 interaction patterns show that models perform well on basic answering tasks, struggle more with repair tasks (especially repeat), have mixed performance on closing sequences, and find complex multi-turn requests most challenging, with Qwen models excelling on the Basic set and Granite models on the RAG set and the Complex Request set. By operationalizing fundamental principles of human conversation, NC-Bench provides a lightweight, extensible, and theory-grounded framework for assessing and improving the conversational abilities of LLMs beyond topical or task-specific benchmarks.

BiasLab: A Multilingual, Dual-Framing Framework for Robust Measurement of Output-Level Bias in Large Language Models

Large Language Models (LLMs) are increasingly deployed in high-stakes contexts where their outputs influence real-world decisions. However, evaluating bias in LLM outputs remains methodologically challenging due to sensitivity to prompt wording, limited multilingual coverage, and the lack of standardized metrics that enable reliable comparison across models. This paper introduces BiasLab, an open-source, model-agnostic evaluation framework for quantifying output-level (extrinsic) bias through a multilingual, robustness-oriented experimental design. BiasLab constructs mirrored probe pairs under a strict dual-framing scheme: an affirmative assertion favoring Target A and a reverse assertion obtained by deterministic target substitution favoring Target B, while preserving identical linguistic structure. To reduce dependence on prompt templates, BiasLab performs repeated evaluation under randomized instructional wrappers and enforces a fixed-choice Likert response format to maximize comparability across models and languages. Responses are normalized into agreement labels using an LLM-based judge, aligned for polarity consistency across framings, and aggregated into quantitative bias indicators with descriptive statistics including effect sizes and neutrality rates. The framework supports evaluation across diverse bias axes, including demographic, cultural, political, and geopolitical topics, and produces reproducible artifacts such as structured reports and comparative visualizations. BiasLab contributes a standardized methodology for cross-lingual and framing-sensitive bias measurement that complements intrinsic and dataset-based audits, enabling researchers and institutions to benchmark robustness and make better-informed deployment decisions.

SentGraph: Hierarchical Sentence Graph for Multi-hop Retrieval-Augmented Question Answering

Traditional Retrieval-Augmented Generation (RAG) effectively supports single-hop question answering with large language models but faces significant limitations in multi-hop question answering tasks, which require combining evidence from multiple documents. Existing chunk-based retrieval often provides irrelevant and logically incoherent context, leading to incomplete evidence chains and incorrect reasoning during answer generation. To address these challenges, we propose SentGraph, a sentence-level graph-based RAG framework that explicitly models fine-grained logical relationships between sentences for multi-hop question answering. Specifically, we construct a hierarchical sentence graph offline by first adapting Rhetorical Structure Theory to distinguish nucleus and satellite sentences, and then organizing them into topic-level subgraphs with cross-document entity bridges. During online retrieval, SentGraph performs graph-guided evidence selection and path expansion to retrieve fine-grained sentence-level evidence. Extensive experiments on four multi-hop question answering benchmarks demonstrate the effectiveness of SentGraph, validating the importance of explicitly modeling sentence-level logical dependencies for multi-hop reasoning.

From Preoperative CT to Postmastoidectomy Mesh Construction:1Mastoidectomy Shape Prediction for Cochlear Implant Surgery

Cochlear Implant (CI) surgery treats severe hearing loss by inserting an electrode array into the cochlea to stimulate the auditory nerve. An important step in this procedure is mastoidectomy, which removes part of the mastoid region of the temporal bone to provide surgical access. Accurate mastoidectomy shape prediction from preoperative imaging improves pre-surgical planning, reduces risks, and enhances surgical outcomes. Despite its importance, there are limited deep-learning-based studies regarding this topic due to the challenges of acquiring ground-truth labels. We address this gap by investigating self-supervised and weakly-supervised learning models to predict the mastoidectomy region without human annotations. We propose a hybrid self-supervised and weakly-supervised learning framework to predict the mastoidectomy region directly from preoperative CT scans, where the mastoid remains intact. Our hybrid method achieves a mean Dice score of 0.72 when predicting the complex and boundary-less mastoidectomy shape, surpassing state-of-the-art approaches and demonstrating strong performance. The method provides groundwork for constructing 3D postmastoidectomy surfaces directly from the corresponding preoperative CT scans. To our knowledge, this is the first work that integrating self-supervised and weakly-supervised learning for mastoidectomy shape prediction, offering a robust and efficient solution for CI surgical planning while leveraging 3D T-distribution loss in weakly-supervised medical imaging.

Overcoming the Retrieval Barrier: Indirect Prompt Injection in the Wild for LLM Systems

Large language models (LLMs) increasingly rely on retrieving information from external corpora. This creates a new attack surface: indirect prompt injection (IPI), where hidden instructions are planted in the corpora and hijack model behavior once retrieved. Previous studies have highlighted this risk but often avoid the hardest step: ensuring that malicious content is actually retrieved. In practice, unoptimized IPI is rarely retrieved under natural queries, which leaves its real-world impact unclear. We address this challenge by decomposing the malicious content into a trigger fragment that guarantees retrieval and an attack fragment that encodes arbitrary attack objectives. Based on this idea, we design an efficient and effective black-box attack algorithm that constructs a compact trigger fragment to guarantee retrieval for any attack fragment. Our attack requires only API access to embedding models, is cost-efficient (as little as $0.21 per target user query on OpenAI's embedding models), and achieves near-100% retrieval across 11 benchmarks and 8 embedding models (including both open-source models and proprietary services). Based on this attack, we present the first end-to-end IPI exploits under natural queries and realistic external corpora, spanning both RAG and agentic systems with diverse attack objectives. These results establish IPI as a practical and severe threat: when a user issued a natural query to summarize emails on frequently asked topics, a single poisoned email was sufficient to coerce GPT-4o into exfiltrating SSH keys with over 80% success in a multi-agent workflow. We further evaluate several defenses and find that they are insufficient to prevent the retrieval of malicious text, highlighting retrieval as a critical open vulnerability.

Using Grok to Avoid Personal Attacks While Correcting Misinformation on X

Correcting misinformation in public online spaces often exposes users to hostility and ad hominem attacks, discouraging participation in corrective discourse. This study presents empirical evidence that invoking Grok, the native large language model on X, rather than directly confronting other users, is associated with different social responses during misinformation correction. Using an observational design, 100 correction replies across five high-conflict misinformation topics were analyzed, with corrections balanced between Grok-mediated and direct human-issued responses. The primary outcome was whether a correction received at least one ad hominem attack within a 24-hour window. Ad hominem attacks occurred in 72 percent of human-issued corrections and in none of the Grok-mediated corrections. A chi-square test confirmed a statistically significant association with a large effect size. These findings suggest that AI-mediated correction may alter the social dynamics of public disagreement by reducing interpersonal hostility during misinformation responses.

Enhancing Multilingual RAG Systems with Debiased Language Preference-Guided Query Fusion

Multilingual Retrieval-Augmented Generation (mRAG) systems often exhibit a perceived preference for high-resource languages, particularly English, resulting in the widespread adoption of English pivoting. While prior studies attribute this advantage to the superior English-centric capabilities of Large Language Models (LLMs), we find that such measurements are significantly distorted by structural priors inherent in evaluation benchmarks. Specifically, we identify exposure bias and a gold availability prior-both driven by the disproportionate concentration of resources in English-as well as cultural priors rooted in topic locality, as factors that hinder accurate assessment of genuine language preference. To address these biases, we propose DeLP (Debiased Language Preference), a calibrated metric designed to explicitly factor out these structural confounds. Our analysis using DeLP reveals that the previously reported English preference is largely a byproduct of evidence distribution rather than an inherent model bias. Instead, we find that retrievers fundamentally favor monolingual alignment between the query and the document language. Building on this insight, we introduce DELTA (DEbiased Language preference-guided Text Augmentation), a lightweight and efficient mRAG framework that strategically leverages monolingual alignment to optimize cross-lingual retrieval and generation. Experimental results demonstrate that DELTA consistently outperforms English pivoting and mRAG baselines across diverse languages.

Comparative Efficiency Analysis of Lightweight Transformer Models: A Multi-Domain Empirical Benchmark for Enterprise NLP Deployment

In the rapidly evolving landscape of enterprise natural language processing (NLP), the demand for efficient, lightweight models capable of handling multi-domain text automation tasks has intensified. This study conducts a comparative analysis of three prominent lightweight Transformer models - DistilBERT, MiniLM, and ALBERT - across three distinct domains: customer sentiment classification, news topic classification, and toxicity and hate speech detection. Utilizing datasets from IMDB, AG News, and the Measuring Hate Speech corpus, we evaluated performance using accuracy-based metrics including accuracy, precision, recall, and F1-score, as well as efficiency metrics such as model size, inference time, throughput, and memory usage. Key findings reveal that no single model dominates all performance dimensions. ALBERT achieves the highest task-specific accuracy in multiple domains, MiniLM excels in inference speed and throughput, and DistilBERT demonstrates the most consistent accuracy across tasks while maintaining competitive efficiency. All results reflect controlled fine-tuning under fixed enterprise-oriented constraints rather than exhaustive hyperparameter optimization. These results highlight trade-offs between accuracy and efficiency, recommending MiniLM for latency-sensitive enterprise applications, DistilBERT for balanced performance, and ALBERT for resource-constrained environments.

The Anatomy of Conversational Scams: A Topic-Based Red Teaming Analysis of Multi-Turn Interactions in LLMs

As LLMs gain persuasive agentic capabilities through extended dialogues, they introduce novel risks in multi-turn conversational scams that single-turn safety evaluations fail to capture. We systematically study these risks using a controlled LLM-to-LLM simulation framework across multi-turn scam scenarios. Evaluating eight state-of-the-art models in English and Chinese, we analyze dialogue outcomes and qualitatively annotate attacker strategies, defensive responses, and failure modes. Results reveal that scam interactions follow recurrent escalation patterns, while defenses employ verification and delay mechanisms. Furthermore, interactional failures frequently stem from safety guardrail activation and role instability. Our findings highlight multi-turn interactional safety as a critical, distinct dimension of LLM behavior.