261 stories tagged LLMs · page 10 of 11
A new arXiv paper quantifies how floating-point precision issues in large language models lead to chaotic behavior. The research highlights the need for better numerical stability in AI systems.
Researchers introduce GeoAgentBench, a dynamic evaluation framework for LLM-based geographic information systems (GIS). It addresses gaps in static testing by assessing real-time, multimodal spatial analysis capabilities.
Researchers introduce a lightweight parallel monitoring system to detect and recover from reasoning degradation in LLM agents. The Cognitive Companion reduces overhead compared to existing solutions. The study shows promise in mitigating issues like looping and drift in multi-step tasks.
The latest installment in the 'LLM from scratch' series covers gradient accumulation, a technique for training large models on limited hardware. This method enables efficient local training of a 32K-parameter model.
A new study on Llama 3.1 8B Instruct shows that the cognitive core of an agent exhibits attractor-like behavior in activation space. This suggests persistent architecture in large language models. The findings could reshape our understanding of LLM cognition.
Researchers propose a method to improve factuality in long-form LLM outputs by calibrating reasoning confidence. This approach reduces overconfident incorrect claims without post-hoc revision.
Researchers introduce a new evaluation method called Filtered Reasoning Score to assess the quality of reasoning in LLMs. This metric highlights that high accuracy doesn't always indicate sound reasoning, as models may rely on memorization or over-optimization.
Researchers introduce HORIZON, a diagnostic benchmark to analyze failures in LLM-based agents on long-horizon tasks. The study highlights the need for better characterization of these failures to improve agentic systems.
Researchers introduce a novel framework to profile tool-using LLM agents in organizational settings. The A-R space measures action rates and refusal signals, offering insights into agent behavior under different autonomy levels.
A new research paper proposes a shift from retrieval-augmented generation to personal wiki-style memory architectures for LLMs. This approach could revolutionize how AI systems retain and utilize knowledge over time.
A new study reveals that large language models, including GPT-4o, perform poorly on abstract meaning comprehension tasks. The findings highlight significant challenges in interpreting non-concrete, high-level semantics.
Researchers propose a method to improve the calibration of large language models (LLMs) without labeled data. The technique leverages the models' internal signals to reduce overconfidence in answers.
Researchers introduce the Cognitive Synergy Framework to improve humor generation in LLMs. The approach uses six cognitive personas to create humorous content, addressing the challenge of incongruity in standard training methods.
Researchers introduce CoSToM, a method to improve LLMs' intrinsic Theory-of-Mind capabilities. The study highlights gaps in current models' ability to generalize social reasoning beyond prompt scaffolding.
Researchers demonstrate that a tutor-student multi-agent system can enhance LLM problem-solving abilities beyond individual capabilities. The approach leverages structured social interaction to achieve synergistic effects in coding tasks.
Researchers propose a new framework to create synthetic physician-physician case discussions using LLMs, addressing privacy concerns. This could enhance AI agents' clinical reasoning capabilities without compromising patient data.
Researchers introduce Sequence-Level PPO (SPPO), a new method to improve reasoning in LLMs by addressing temporal credit assignment and memory costs. SPPO enhances stability and efficiency over traditional PPO approaches.
Researchers introduce SEA-Eval, a benchmark to evaluate self-evolving agents that can learn and adapt across tasks. This addresses limitations of current episodic LLM-based agents.
Researchers introduce PilotBench, a benchmark for testing LLMs on flight trajectory and attitude prediction with safety constraints. The dataset includes 708 real-world general aviation trajectories with synchronized telemetry data.
A new survey and benchmark reveal that while large language models excel at medical exams, their clinical reasoning falls short. The study underscores the need for robust medical reasoning capabilities in real-world healthcare settings.
Researchers introduce Springdrift, a persistent runtime for LLM agents that prioritizes auditing. This could revolutionize how we track and verify AI agent actions.
Researchers introduce SepSeq, a training-free framework that inserts separator tokens to solve attention dispersion in LLMs processing long numerical data. This plug-and-play method significantly boosts performance without requiring model retraining or architectural changes.
A new study highlights the risks of malicious attacks on the LLM supply chain, demonstrating how agents can be compromised. The findings underscore the need for robust security measures in AI development.
Qualixar OS emerges as the first application-layer operating system designed for universal AI agent orchestration, supporting 10 LLM providers and 8+ frameworks. It introduces execution semantics for 12 distinct multi-agent topologies and a novel LLM-driven design engine called Forge.