Daisy Christodoulou (@daisychristodoulou): "LLMs hallucinate. How many hallucinations is too many? Google & Eedi have developed an LLM tutor with just a 0.14% error rate. This is incredibly impressive - but even so may still cause problems."

The app for independent voices

Daisy Christodoulou

No More Marking

LLMs hallucinate.

How many hallucinations is too many?

Google & Eedi have developed an LLM tutor with just a 0.14% error rate. This is incredibly impressive - but even so may still cause problems.

Also, whilst one in a thousand errors sounds low, it’s possible that one lesson’s worth of conversation with the chatbot might include 50 or so messages, which would effectively 50x the error rate.

Daisy Christodoulou

Jan 12

at

9:03 PM

#nojs-banner { position: fixed; bottom: 0; left: 0; padding: 16px 16px 16px 32px; width: 100%; box-sizing: border-box; background: red; color: white; font-family: -apple-system, "Segoe UI", Roboto, Helvetica, Arial, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol"; font-size: 13px; line-height: 13px; } #nojs-banner a { color: inherit; text-decoration: underline; } This site requires JavaScript to run correctly. Please turn on JavaScript or unblock scripts