Pages 5 and 6 illustrating examples of iterative reflective equilibrium from a language model.

Formal explication of reflective equilibrium and technical pipeline (pages 8 and 10).

Testing results (Table 3) and cross-version comparisons (Table 4) from pages 16 and 17.

Mitigation results (Table 5) and external LLM comparisons (Table 6) with agreement among ChatGLM, Vicuna, and GPT-3.5-Turbo (Figure 6) on pages 18 and 19.

Reflecting on our intuitions and principles until they are logically consistent is hard. Can #AI do it?

Ma et al. explicate #ReflectiveEquilibrium (RE) and test how #LLMs iteratively achieve RE on moral scenarios from the #ETHICS benchmark.

doi.org/10.1145/3722554

#xPhi #xJur

This Wednesday, October 13, I give a talk on "Is #ReflectiveEquilibrium Really a Method?" online at the EXTRA.14 Colloquium. I am especially exctited to present my ideas to this group, as it it was some talks in last term's EXTRA.13 series that inspired me to return to this project 😀🚀 #philsky

Cover

Contents (1 of 2)

Contents (2 of 2)

Pages 20 and 21, featuring a schematic overview of reflective equilibrium (Figure 2.1)

🚨 Free book alert! (ePub + PDF!)

This book reveals each step of #reflectiveEquilibrium (RE), so anyone can use it.

In the process, the book reveals both the benefits and the limits of RE.

Get it at doi.org/10.1007/978-...

#Logic #Ethics #Politics #Science #Medicine #Epistemology #Metaphilosophy

Can #children learn morality via #reflectiveEquilibrium?
@John_Tillson argues that they can: http://buff.ly/2tARCyO
#philosophy #P4C