Positive users hail proper calibration of LLM forecasters via the Goodfire-EternisAI partnership as the holy grail for fixing overconfidence, while the negative reply dismisses the announcement with hostile demands for moneymaking details.
Based on 3 visible X reactions from 9 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@GoodfireAI @EternisAI Proper calibration is the holy grail of LLM forecasting. Standard prompting completely fails to fix overconfidence because models just sound confidently wrong. Using representation probes to extract and adjust internal uncertainty mechanics is the right path forward. 📈
@GoodfireAI @EternisAI Ok well who is it bitch. Im tryna make some money
@GoodfireAI @EternisAI LLM forecasters will put those astrologers into dustl.
the probe said england 100% confidence to win this year btw https://twitter.com/GoodfireAI/status/2075611273061175479
Positive users hail proper calibration of LLM forecasters via the Goodfire-EternisAI partnership as the holy grail for fixing overconfidence, while the negative reply dismisses the announcement with hostile demands for moneymaking details.
Based on 3 visible X reactions from 9 accounts; directional sample.
Ask a question below.
Published answers will appear here.