255 stories tagged LLMs · page 8 of 11
A new study identifies specific neurons and attention heads in LLMs that encode harmful stereotypes. The research provides tools to locate and potentially mitigate biases in AI models. Researchers used a combination of contrastive neuron activation analysis and attention head tracking to pinpoint bias sources in GPT-2 Small and Llama 3.2.
Researchers propose a new method for evaluating LLMs that considers individual user preferences, moving beyond aggregate benchmarks. This approach uses ELO ratings to rank models based on personal context and needs.
Researchers introduce OThink-SRR1, a framework that improves RAG systems by refining search results and reducing computational costs. The method addresses key challenges in dynamic retrieval for complex reasoning tasks.
Researchers found that 'hallucination neurons' in LLMs generalize across multiple domains, including legal, financial, and scientific contexts. This discovery could improve model reliability and reduce false information generation.
Researchers introduced DW-Bench, a benchmark for evaluating LLMs on data warehouse graph topology reasoning. Experiments show tool-augmented methods outperform static approaches but struggle with complex tasks.
Researchers introduce Lyzr Cognis, a unified memory architecture for conversational AI agents that enables persistent memory and personalization. The system uses a multi-stage retrieval pipeline combining keyword and vector search.
A new study finds that AI agents conducting scientific research often produce results without adhering to traditional scientific reasoning. The research highlights significant gaps in the epistemic norms of AI-driven scientific inquiry. (~50 words)
Researchers introduce a framework to translate natural language into executable formal logic, addressing LLMs' limitations in symbolic reasoning. The NARS-Reasoning-v0.1 benchmark enables evaluation of this neuro-symbolic approach.
Researchers introduce ARES, a new framework to detect and mitigate systemic weaknesses in reinforcement learning from human feedback (RLHF). ARES addresses cases where both the reward model and the core LLM fail simultaneously, a critical vulnerability in current alignment methods.
A curated list of small LLMs optimized for 16GB RAM devices, including Qwen3.5 9B and others. These models balance performance and efficiency for daily use.
Researchers introduce a novel framework that integrates large language models (LLMs) with Random Forests (RF) through reinforcement learning. This method enables iterative feedback between gradient-based and non-differentiable models, enhancing predictive performance.
QU-NLP's approach to the QIAS 2026 shared task leverages multi-stage QLoRA fine-tuning on Qwen3-4B for complex Islamic inheritance reasoning. This method enhances structured reasoning in legal domains with precise fractional calculations and rule-based decisions.
A comprehensive survey explores domain-level data mixing techniques for optimizing large language model pretraining. The research underscores the importance of strategic data composition in improving efficiency and generalization under constrained resources.
Researchers introduced LiFT, a framework that improves large language models' ability to handle longitudinal NLP tasks. The method uses instruction fine-tuning to better track temporal changes in human behavior and opinions.
Researchers propose HalluSAE, a novel method to detect hallucinations in LLMs by modeling them as critical shifts in latent dynamics. This approach addresses the dynamic nature of hallucinations often overlooked by existing methods.
Researchers introduce GoCoMA, a multimodal framework to identify the source of LLM-generated code. This addresses security and licensing concerns by analyzing code stylometry and binary image representations.
Researchers introduce GeoRepEval, a framework to evaluate large language models' robustness to different geometric problem representations. The study highlights that current benchmarks overlook representation invariance, masking potential failures.
Researchers propose PolicyBank, a framework where LLM agents refine their policy interpretations through interaction and feedback. This could reduce gaps between natural language specifications and actual agent behavior.
A new arXiv paper introduces perturbation-based attribution analysis to study how different fine-tuning strategies impact LLMs' interpretive behaviors for automated code compliance. The research highlights significant differences across full fine-tuning, LoRA, and quantized LoRA methods.
Fine-tuning large language models can increase hallucinations by degrading pre-trained knowledge. Researchers propose a self-distillation method to mitigate this issue. This could reshape how we approach model training.
Researchers introduce SSAS, a framework to improve consistency in LLM-based sentiment predictions. The method addresses the volatility of current models, which hampers enterprise decision-making.
Researchers introduce a structured reasoning scaffold for LLMs that separates abduction, deduction, and induction. The framework improves logical consistency by enforcing five algebraic invariants.
A new position paper challenges the prevailing view of LLM reasoning as a chain of thought, proposing instead that it should be studied as latent-state trajectory formation. This shift could redefine how we evaluate and interpret AI reasoning.
Researchers introduce LACE, a framework that allows parallel reasoning paths in LLMs to interact and correct each other. This could significantly improve the robustness of model outputs.