When Better Perplexity Hides Less Variety
SOL reveals distributional failures that perplexity and token entropy can miss, offering evaluation teams a new way to compare generated text.
SOL reveals distributional failures that perplexity and token entropy can miss, offering evaluation teams a new way to compare generated text.
PaperBenchX shows why scientific agents must regenerate and validate evidence, not merely execute specialized software.
A mechanistic audit of 25 language models shows why distinguishable scoring criteria may still lack independent control over an AI evaluator's decisions.
How cross-model failure evidence can improve shared AI agent harnesses, and when the development-cost advantage holds.
Decision-guided world modeling improves language-agent training by selecting which environmental information deserves prediction supervision.
MultINav shows why robots must verify changes in their environment and distinguish successful physical interactions from completed navigation tasks.
A suicide-risk AI study shows how complementary factor verification can improve explanations while upstream classification errors can suppress relevant evidence.
GeoReform shows how improving the structured information supplied to a multimodal model can improve geometric reasoning without changing the solver.
How rendering untrusted text as images can reduce prompt injection—and why model behavior, transcription fidelity, and costs determine whether the defense is deployable.
REVEAL shows how successful and failed frontier-agent runs can produce activation controls for smaller models, with promising benchmark gains but untested generalization.