LLM Research
New research introduces Accordion-Thinking, a self-regulated approach that compresses reasoning steps dynamically to improve LLM efficiency while maintaining readable chain-of-thought outputs.
LLM
Researchers introduce a unified benchmark for evaluating multi-agent LLM frameworks, providing systematic analysis of how autonomous AI agents collaborate on complex tasks.
LLM Efficiency
New research proposes dynamic precision routing to optimize computational resources across multi-step LLM interactions, balancing quality and efficiency through adaptive quantization strategies.
AI Agents
New research introduces MARS, a modular agent with reflective search capabilities designed to automate AI research tasks through intelligent decomposition and self-correction.
LLM Interpretability
New research presents evidence that LLM self-explanations can help predict model behavior, offering a positive case for faithfulness in AI interpretability.
LLM Security
Researchers detail how prompt injection, jailbreaking, and gradient-based attacks systematically defeat the layered safety mechanisms designed to keep large language models aligned and secure.
deepfake detection
As deepfake cyberattacks grow more sophisticated, security experts argue that traditional perimeter defenses are insufficient. A trust-centric approach may offer better protection against AI-generated threats.
AI Hardware
Startup Positron secures $230M Series B to develop alternative AI chips, potentially reshaping the hardware landscape that powers video generation and synthetic media systems.
LLM evaluation
New research proposes PeerRank, a system where LLMs evaluate each other through web-grounded peer review with built-in bias controls, potentially transforming how we benchmark AI models.
deepfakes
An AI security company's hiring process became a real-world test of deepfake detection when a synthetic candidate attempted to infiltrate through video interviews.
deepfake detection
Pindrop integrates its real-time fraud and deepfake defense technology with NICE's CXone and CX AI platform, bringing voice authentication to enterprise contact centers.
AI Security
Researchers reveal how imperceptible visual perturbations embedded in images can hijack vision-language models, bypassing safety filters and manipulating AI outputs without human detection.