5 stories tagged Detection
Scientists have discovered that AI hallucinations are easier to detect in intermediate layers of AI models, not just the final output. They've developed a method to automatically find the best layers for spotting these errors, making AI more reliable.
Researchers developed DetectRL-X, a benchmark to test AI text detectors in real-world, multilingual scenarios. This tool evaluates detectors across 8 languages, aiming to improve reliability and governance of AI-generated content.
Researchers found a way to bypass watermarks on AI-generated text, raising concerns about detecting fake content. This could make it harder to tell if text was written by a human or an AI.
Researchers have developed a new way to watermark AI-generated text without reducing its quality. SLAM marks text structure rather than word choices, making it harder to detect and remove.
Researchers propose a weak supervision framework to detect hallucinations in large language models. This approach enables hallucination detection from internal activations alone at inference time.