28 articles tagged with #semantic-analysis. AI-curated summaries with sentiment analysis and key takeaways from 50+ sources.
Researchers propose the Adversarial Prompt Disentanglement (APD) framework, a defense mechanism that identifies and neutralizes malicious components in LLM inputs before processing. The system combines semantic decomposition, graph-based intent classification, and transformer-based detection to reduce harmful outputs by over 85% while maintaining model performance.
Researchers present a hybrid neuro-symbolic architecture that combines formal logic with neural semantic analysis to verify LLM outputs in high-stakes domains like healthcare. The system achieves over 83% hallucination detection rates for structured data and 72% for semantic fabrications while reducing report creation time by 30%, demonstrating practical safeguards for deploying LLMs in data-sensitive applications.
A new arXiv study challenges the assumption that Chain of Thought reasoning traces in large language models reflect genuine internal reasoning processes. Researchers found that models trained on corrupted, semantically meaningless intermediate steps perform comparably to those trained on correct reasoning traces, suggesting that intermediate tokens function more as statistical patterns than transparent reasoning proxies.
Researchers found that large language models frequently arrive at correct code predictions through flawed reasoning, with performance dropping up to 70% when code undergoes semantics-preserving mutations. The study reveals substantial gaps between apparent accuracy and genuine semantic understanding, questioning the reliability of LLMs for critical programming tasks.
Researchers introduce Distributional Semantics Tracing (DST), a new framework for explaining hallucinations in large language models by tracking how semantic representations drift across neural network layers. The method reveals that hallucinations occur when models are pulled toward contextually inconsistent concepts based on training correlations rather than actual prompt context.
Researchers propose SemKey, a novel framework that addresses key limitations in EEG-to-text decoding by preventing hallucinations and improving semantic fidelity through decoupled guidance objectives. The system redesigns neural encoder-LLM interaction and introduces new evaluation metrics beyond BLEU scores to achieve state-of-the-art performance in brain-computer interfaces.
Researchers developed LLM-MLFFN, a new framework combining large language models with multi-level feature fusion to classify autonomous vehicle driving behaviors. The system achieves over 94% accuracy on the Waymo dataset by integrating numerical driving data with semantic features extracted through LLMs.
Researchers comparing hybrid language models (mixing attention and recurrent layers) against pure transformers using Olmo weights find that hybrids excel at semantic state tracking but underperform on syntactic tasks like bracket matching. The analysis reveals that recurrent layers and attention mechanisms have complementary strengths, with gains concentrated in open-class words and semantic tasks rather than function words or n-gram prediction.
Researchers introduced the Meaning Intelligence Framework (MIF), a nine-dimension evaluation schema that improves AI systems' ability to understand Nigerian public discourse by separating surface sentiment from true communicative intent. The framework increased register classification accuracy from 33.3% to 73.3% when applied to frontier language models, revealing that context failure—not translation failure—is the primary limitation of current AI systems on Nigerian languages.
Researchers developed a semantic-timescale analysis pipeline to compare how human and AI-generated speech organize semantic content over time. Using autocorrelation measures on word specificity and contextual similarity, they found that temporal clustering of generic versus specific vocabulary distinguishes human narratives from LLM outputs, revealing non-trivial structural differences beyond static word frequency.
Researchers conducted a large-scale semantic analysis of 8,954 definitions and 2,700 scale items across 14,000+ publications to map how learner agency and autonomy are conceptualized and measured. They identified three core dimensions (task regulation, intrinsic motivation, and sociocultural action) and found that existing measurement scales systematically underrepresent the sociocultural aspect, while current generative AI applications in education narrowly focus on learning control.
Researchers introduce SPG-LLM, a novel approach that leverages large language models to optimize the grounding process in classical planning by identifying irrelevant objects and actions before computation. The method achieves significantly faster grounding times—often by orders of magnitude—across seven challenging benchmarks while maintaining or improving plan quality.
Researchers have developed a new method called Semantic Correlation Descriptors (SCDs) to identify whether a specific dataset was used to train a machine learning model by analyzing the spurious correlations embedded in its learned structure. This white-box approach outperforms existing black-box membership inference techniques, achieving up to 60% higher accuracy in detecting dataset membership across natural language and medical text classification tasks.
Researchers introduce GPF-LiveNews, a streaming evaluation protocol that audits how large language models frame news differently based on group identities and prompts. Testing 23 models across 42 identity labels reveals that policy-oriented prompts trigger stronger semantic shifts in framing, while sentiment variation remains inconsistent, highlighting the need for continuous monitoring of LLM outputs in production environments.
Researchers present a modular LLM-based architecture for detecting and quantifying human values in text, addressing the need for ethical decision-making in autonomous AI systems. The approach separates value conceptualization from detection, enabling scalable application across different ethical frameworks and demonstrating strong performance on the ValueEval dataset.
Semia is a static auditor for LLM-driven agent skills that uses constraint-guided synthesis to analyze security risks in hybrid code-and-prose configurations. Testing 13,728 real-world skills from public marketplaces, Semia identified critical semantic vulnerabilities in over half and achieved 97.7% recall, significantly outperforming existing security tools.
Researchers propose a novel framework treating Large Language Models as attention-informed Neural Topic Models, enabling interpretable topic extraction from documents. The approach combines white-box interpretability analysis with black-box long-context LLM capabilities, demonstrating competitive performance on topic modeling tasks while maintaining semantic clarity.
Researchers introduce Agent Mentor, an open-source analytics pipeline that monitors and automatically improves AI agent behavior by analyzing execution logs and iteratively refining system prompts with corrective instructions. The framework addresses variability in large language model-based agent performance caused by ambiguous prompt formulations, demonstrating consistent accuracy improvements across multiple configurations.
Researchers introduce TagCC, a novel deep clustering framework that combines Large Language Models with contrastive learning to enhance tabular data analysis by incorporating semantic knowledge from feature names and values. The approach bridges the gap between statistical co-occurrence patterns and intrinsic semantic understanding, demonstrating significant performance improvements over existing methods in finance and healthcare applications.
Researchers introduce ICR (Inductive Conceptual Rating), a new qualitative metric for evaluating meaning in large language model text summaries that goes beyond simple word similarity. The study found that while LLMs achieve high linguistic similarity to human outputs, they significantly underperform in semantic accuracy and capturing contextual meanings.
Researchers have developed FinTexTS, a new large-scale dataset that pairs financial news with stock price data using semantic matching and multi-level categorization. The framework uses embedding-based matching and LLMs to classify news into four levels (macro, sector, related company, and target company) for improved stock price forecasting accuracy.
Researchers introduce Latent Self-Consistency (LSC), a new method for improving Large Language Model output reliability across both short and long-form reasoning tasks. LSC uses learnable token embeddings to select semantically consistent responses with only 0.9% computational overhead, outperforming existing consistency methods like Self-Consistency and Universal Self-Consistency.
Researchers have developed Taxoria, a new taxonomy enrichment pipeline that uses Large Language Models to enhance existing taxonomies by proposing, validating, and integrating new nodes. The system addresses limitations in current taxonomies such as limited coverage and outdated information while including hallucination mitigation and provenance tracking.
Researchers introduce Temporal Sparse Autoencoders (T-SAEs), a new method that improves AI model interpretability by incorporating temporal structure of language through contrastive loss. The technique enables better separation of semantic from syntactic features and recovers smoother, more coherent semantic concepts without sacrificing reconstruction quality.
Researchers developed an automated framework to evaluate Large Language Models' effectiveness in translating Mandarin Chinese to English, comparing GPT-4, GPT-4o, and DeepSeek against Google Translate. While LLMs performed well on news translation, they showed varying results with literary texts, with DeepSeek excelling at cultural subtleties and GPT-4o/DeepSeek better at semantic conservation.
