LLLMs: A Data-Driven Survey of Evolving Research on Limitations of Large Language Models

Large language model (LLM) research has grown rapidly, along with increasing concern about their limitations such as failures in reasoning, hallucinations, and limited multilingual capability. In this survey, we conduct a data-driven, semi-automated review of research on limitations of LLM (LLLMs) from 2022 to 2024 using a bottom-up approach. From a corpus of 250,000 ACL and arXiv papers, we identify 14,648 relevant papers using keyword filtering, LLM-based classification, validated against expert labels, and topic clustering (via two approaches, HDBSCAN+BERTopic and LlooM). We find that LLM-related research increases over fivefold in ACL and fourfold in arXiv. Since 2022, LLLMs research grows even faster, reaching over 30% of LLM papers by late 2024. Reasoning remains the most studied limitation, followed by generalization, hallucination, bias, and security. The distribution of topics in the ACL dataset stays relatively stable over time, while arXiv shifts toward safety and controllability (with topics like security risks, alignment, hallucinations, knowledge editing), and multimodality between 2022 and 2024. We release a dataset of annotated abstracts and a validated methodology, and offer a quantitative view of trends in LLM limitations research.

B-score: Detecting biases in large language models using response history

Large language models (LLMs) often exhibit strong biases, e.g, against women or in favor of the number 7. We investigate whether LLMs would be able to output less biased answers when allowed to observe their prior answers to the same question in a multi-turn conversation. To understand which types of questions invite more biased answers, we test LLMs on our proposed set of questions that span 9 topics and belong to three types: (1) Subjective; (2) Random; and (3) Objective. Interestingly, LLMs are able to "de-bias" themselves in a multi-turn conversation in response to questions that seek an Random, unbiased answer. Furthermore, we propose B-score, a novel metric that is effective in detecting biases to Subjective, Random, Easy, and Hard questions. On MMLU, HLE, and CSQA, leveraging B-score substantially improves the verification accuracy of LLM answers (i.e, accepting LLM correct answers and rejecting incorrect ones) compared to using verbalized confidence scores or the frequency of single-turn answers alone. Code and data are available at: https://b-score.github.io.

Online Iterative Self-Alignment for Radiology Report Generation

Radiology Report Generation (RRG) is an important research topic for relieving radiologist' heavy workload. Existing RRG models mainly rely on supervised fine-tuning (SFT) based on different model architectures using data pairs of radiological images and corresponding radiologist-annotated reports. Recent research has shifted focus to post-training improvements, aligning RRG model outputs with human preferences using reinforcement learning (RL). However, the limited data coverage of high-quality annotated data poses risks of overfitting and generalization. This paper proposes a novel Online Iterative Self-Alignment (OISA) method for RRG that consists of four stages: self-generation of diverse data, self-evaluation for multi-objective preference data,self-alignment for multi-objective optimization and self-iteration for further improvement. Our approach allows for generating varied reports tailored to specific clinical objectives, enhancing the overall performance of the RRG model iteratively. Unlike existing methods, our frame-work significantly increases data quality and optimizes performance through iterative multi-objective optimization. Experimental results demonstrate that our method surpasses previous approaches, achieving state-of-the-art performance across multiple evaluation metrics.

An Empirical Study of OpenAI API Discussions on Stack Overflow

The rapid advancement of large language models (LLMs), represented by OpenAI's GPT series, has significantly impacted various domains such as natural language processing, software development, education, healthcare, finance, and scientific research. However, OpenAI APIs introduce unique challenges that differ from traditional APIs, such as the complexities of prompt engineering, token-based cost management, non-deterministic outputs, and operation as black boxes. To the best of our knowledge, the challenges developers encounter when using OpenAI APIs have not been explored in previous empirical studies. To fill this gap, we conduct the first comprehensive empirical study by analyzing 2,874 OpenAI API-related discussions from the popular Q&A forum Stack Overflow. We first examine the popularity and difficulty of these posts. After manually categorizing them into nine OpenAI API-related categories, we identify specific challenges associated with each category through topic modeling analysis. Based on our empirical findings, we finally propose actionable implications for developers, LLM vendors, and researchers.

Exploring the generalization of LLM truth directions on conversational formats

Several recent works argue that LLMs have a universal truth direction where true and false statements are linearly separable in the activation space of the model. It has been demonstrated that linear probes trained on a single hidden state of the model already generalize across a range of topics and might even be used for lie detection in LLM conversations. In this work we explore how this truth direction generalizes between various conversational formats. We find good generalization between short conversations that end on a lie, but poor generalization to longer formats where the lie appears earlier in the input prompt. We propose a solution that significantly improves this type of generalization by adding a fixed key phrase at the end of each conversation. Our results highlight the challenges towards reliable LLM lie detectors that generalize to new settings.

From Generation to Detection: A Multimodal Multi-Task Dataset for Benchmarking Health Misinformation

Infodemics and health misinformation have significant negative impact on individuals and society, exacerbating confusion and increasing hesitancy in adopting recommended health measures. Recent advancements in generative AI, capable of producing realistic, human like text and images, have significantly accelerated the spread and expanded the reach of health misinformation, resulting in an alarming surge in its dissemination. To combat the infodemics, most existing work has focused on developing misinformation datasets from social media and fact checking platforms, but has faced limitations in topical coverage, inclusion of AI generation, and accessibility of raw content. To address these issues, we present MM Health, a large scale multimodal misinformation dataset in the health domain consisting of 34,746 news article encompassing both textual and visual information. MM Health includes human-generated multimodal information (5,776 articles) and AI generated multimodal information (28,880 articles) from various SOTA generative AI models. Additionally, We benchmarked our dataset against three tasks (reliability checks, originality checks, and fine-grained AI detection) demonstrating that existing SOTA models struggle to accurately distinguish the reliability and origin of information. Our dataset aims to support the development of misinformation detection across various health scenarios, facilitating the detection of human and machine generated content at multimodal levels.

Hydra: Structured Cross-Source Enhanced Large Language Model Reasoning

Retrieval-augmented generation (RAG) enhances large language models (LLMs) by incorporating external knowledge. Current hybrid RAG system retrieves evidence from both knowledge graphs (KGs) and text documents to support LLM reasoning. However, it faces challenges like handling multi-hop reasoning, multi-entity questions, multi-source verification, and effective graph utilization. To address these limitations, we present Hydra, a training-free framework that unifies graph topology, document semantics, and source reliability to support deep, faithful reasoning in LLMs. Hydra handles multi-hop and multi-entity problems through agent-driven exploration that combines structured and unstructured retrieval, increasing both diversity and precision of evidence. To tackle multi-source verification, Hydra uses a tri-factor cross-source verification (source trustworthiness assessment, cross-source corroboration, and entity-path alignment), to balance topic relevance with cross-modal agreement. By leveraging graph structure, Hydra fuses heterogeneous sources, guides efficient exploration, and prunes noise early. Comprehensive experiments on seven benchmark datasets show that Hydra achieves overall state-of-the-art results on all benchmarks with GPT-3.5, outperforming the strong hybrid baseline ToG-2 by an average of 20.3% and up to 30.1%. Furthermore, Hydra enables smaller models (e.g., Llama-3.1-8B) to achieve reasoning performance comparable to that of GPT-4-Turbo.

ToxicTone: A Mandarin Audio Dataset Annotated for Toxicity and Toxic Utterance Tonality

Despite extensive research on toxic speech detection in text, a critical gap remains in handling spoken Mandarin audio. The lack of annotated datasets that capture the unique prosodic cues and culturally specific expressions in Mandarin leaves spoken toxicity underexplored. To address this, we introduce ToxicTone -- the largest public dataset of its kind -- featuring detailed annotations that distinguish both forms of toxicity (e.g., profanity, bullying) and sources of toxicity (e.g., anger, sarcasm, dismissiveness). Our data, sourced from diverse real-world audio and organized into 13 topical categories, mirrors authentic communication scenarios. We also propose a multimodal detection framework that integrates acoustic, linguistic, and emotional features using state-of-the-art speech and emotion encoders. Extensive experiments show our approach outperforms text-only and baseline models, underscoring the essential role of speech-specific cues in revealing hidden toxic expressions.

BCause: Human-AI collaboration to improve hybrid mapping and ideation in argumentation-grounded deliberation

Public deliberation, as in open discussion of issues of public concern, often suffers from scattered and shallow discourse, poor sensemaking, and a disconnect from actionable policy outcomes. This paper introduces BCause, a discussion system leveraging generative AI and human-machine collaboration to transform unstructured dialogue around public issues (such as urban living, policy changes, and current socio-economic transformations) into structured, actionable democratic processes. We present three innovations: (i) importing and transforming unstructured transcripts into argumentative discussions, (ii) geo-deliberated problem-sensing via a Telegram bot for local issue reporting, and (iii) smart reporting with customizable widgets (e.g., summaries, topic modelling, policy recommendations, clustered arguments). The system's human-AI partnership preserves critical human participation to ensure ethical oversight, contextual relevance, and creative synthesis.

MedScore: Factuality Evaluation of Free-Form Medical Answers

While Large Language Models (LLMs) can generate fluent and convincing responses, they are not necessarily correct. This is especially apparent in the popular decompose-then-verify factuality evaluation pipeline, where LLMs evaluate generations by decomposing the generations into individual, valid claims. Factuality evaluation is especially important for medical answers, since incorrect medical information could seriously harm the patient. However, existing factuality systems are a poor match for the medical domain, as they are typically only evaluated on objective, entity-centric, formulaic texts such as biographies and historical topics. This differs from condition-dependent, conversational, hypothetical, sentence-structure diverse, and subjective medical answers, which makes decomposition into valid facts challenging. We propose MedScore, a new approach to decomposing medical answers into condition-aware valid facts. Our method extracts up to three times more valid facts than existing methods, reducing hallucination and vague references, and retaining condition-dependency in facts. The resulting factuality score significantly varies by decomposition method, verification corpus, and used backbone LLM, highlighting the importance of customizing each step for reliable factuality evaluation.