AI ALIGNMENT FORUM
AF

3204

Honesty

Edited by Yoav Ravid, Multicore last updated 3rd Mar 2021

Honesty means telling the truth and not being deceptive.

External Links:
Against Lie Inflation by Scott Alexander

Related Pages: Meta-Honesty, Deception.

2

2

Posts tagged Honesty

35Truthful LMs as a warm-up for aligned AGI

4y

10

33Paper: Teaching GPT3 to express uncertainty in words

3y

0

29How "honest" is GPT-3?

abramdemski, gwern

5y

4

18How do new models from OpenAI, DeepMind and Anthropic perform on TruthfulQA?

4y

1

-21Lying is Cowardice, not Strategy

Connor Leahy, Gabriel Alfour

2y

21

93How "Discovering Latent Knowledge in Language Models Without Supervision" Fits Into a Broader Alignment Scheme

3y

20

53Contrast Pairs Drive the Empirical Performance of Contrast Consistent Search (CCS)

2y

1

23Truthful AI: Developing and governing AI that does not lie

Owain_Evans, owencb, Lukas Finnveden

4y

9

18Tall Tales at Different Scales: Evaluating Scaling Trends For Deception In Language Models

Felix Hofstätter, Francis Rhys Ward, HarrietW, LAThomson, Ollie J, Patrik Bartak, Sam F. Brown

2y

0

14Truth is Universal: Robust Detection of Lies in LLMs

Lennart Buerger

1y

2

3Ground-Truth Label Imbalance Impairs the Performance of Contrast-Consistent Search (and Other Contrast-Pair-Based Unsupervised Methods)

Tom Angsten, Ami Hays

2y

0

Add Posts