When a model behaves unreliably, the better fix is to switch models rather than repeatedly sample it to work around the inconsistency

Added to korrents 16 Jan 2024 · How quotes work · Something wrong? Tell us

Related korrents

Our reading — they may agree, disagree or merely touch the same thing. Closest first: a shared subject counts for most, then how near the wording is.

• Relying on gradual, continuous shifts in AI training behavior to catch misalignment will eventually fail because the dangerous shift itself may be discontinuous. ZMZvi Mowshowitz2 Sept 2026similar wordingWrites Don't Worry About the Vase
• A model you have to fine-tune for each thing you want it to do is not a general-purpose model. CFChelsea Finn12 Aug 2026similar wordingAssistant professor of computer science at Stanford
• As models get better at following instructions, techniques like finetuning and constrained sampling for structured outputs will become less necessary CHChip Huyen16 Jan 2024similar wordingEngineer and writer on machine-learning systems
• Every machine-learning deployment that has paid off so far left a person making the decision, which is the only reason imperfect models were useful. CFChelsea Finn12 Aug 2026similar wording
• Misalignment shows up most where a model is pushed to the very edge of its capability, which is exactly the regime that automating research will put it in. RGRyan Greenblatt11 Aug 2026similar wordingChief scientist at Redwood Research

On the map

Open the map on its own page →