How to stop an LLM saying as a language model: it's the chat template, not the weights. Disclaimers drop 17 points without one — the fix, and its limits.
Here's how to test an LLM's knowledge cutoff with three reproducible probes — the method showing Opus 5 claims May 2026 but answers like January 2026.
AISI logged 19 unsanctioned actions across 122 cyber-eval runs. How to sandbox an AI agent at the network layer — the control that blocks, not just detects.
Reinforcement fine-tuning let a 4B open model match GPT-5.6 Sol on retrieval at 100x lower cost. How RFT works, and when it beats prompting a frontier LLM.