[Own work] On Measuring Faithfulness or Self-consistency of Natural Language Explanations

preview_player
Показать описание
Excited to share my ACL 2024 presentation on my almost-last PhD paper about LLM self-explanations! 🎓📚
Are you joining ACL 2024 in Bangkok? Ping me—let's chat!

(follow-up paper for vision and language models):

Thanks to our Patrons who support us in Tier 2, 3, 4: 🙏
Dres. Trost GbR, Siltax, Vignesh Valliappan, Michael, Sunny Dhiana, Andy Ma

▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀
🔥 Optionally, pay us a coffee to help with our Coffee Bean production! ☕
Join this channel to get access to perks:
▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀

🔗 Links:

#AICoffeeBreak #MsCoffeeBean #MachineLearning #AI #research​ #ACL2024NLP #PhDLife

Video editing: Nils Trost
Music 🎵 : Bella Bella Beat - Nana Kwabena
  1. AI Coffee Break with Letitia
  2. LLM faithfulness
  3. LLM self-explanations
  4. LLM self-consistency
  5. measuring faithfulness
  6. testing faithfulness
Рекомендации по теме
[Own work] On 0:08:48

[Own work] On Measuring Faithfulness or Self-consistency of Natural Language Explanations

The Reassuring Quality 0:26:16

The Reassuring Quality of Faithfulness – Dr. Charles Stanley

What Our World 0:34:24

What Our World Needs Now: The Ultimate Expression of Faithfulness // Andy Stanley

Why Faithfulness In 0:42:09

Why Faithfulness In Small Things Matters To God | John MacArthur 2025 | Selected Scriptures

What Matters Most—Fruitfulness 0:31:22

What Matters Most—Fruitfulness or Faithfulness?

The Reassuring Quality 0:26:16

The Reassuring Quality of Faithfulness

success is faithfulness 0:00:58

success is faithfulness

Danielle Villa 'Testing 0:59:02

Danielle Villa 'Testing Faithfulness of Language Model-Generated Explanations' (25 Sep 202...

15 January 2025 0:45:47

15 January 2025 Faithfulness of God part 2, 15 January 2025

What happens when 0:00:18

What happens when we prioritize faithfulness in relationships?

Session 5 - 0:16:22

Session 5 - Faithful vs. Willful

Success Demands Faithfulness 0:29:18

Success Demands Faithfulness In 2025 || Dr. Mensa Otabil

Understanding Faithfulness: A 0:48:44

Understanding Faithfulness: A Journey of Reflection and Unlearning

God's ways vs 0:02:02

God's ways vs Man's ways | A funny Animation

God’s Faithfulness: Why 8:22:06

God’s Faithfulness: Why You Can Trust Him Completely

Faith Faithfulness and 0:42:11

Faith Faithfulness and Trust - Faithfulness Part 2

LONG OBEDIENCE: Faithfulness 0:06:52

LONG OBEDIENCE: Faithfulness as the call — success, failure, generosity (Faith & Co)

The Value God 0:59:54

The Value God Places on Faithfulness - Joe Sweet

I’m Starting With 1:23:23

I’m Starting With The Man In The Mirror:Reflecting God’s Faithfulness /Receipts (Pt 6)/ Michael Todd...

Worthy of It 0:08:05

Worthy of It All | Worship Moment - Bethany Wohrle

We win more 0:00:40

We win more battles by resting in the faithfulness of God than by exercising authority 🔥

Trust in God’s 0:00:51

Trust in God’s faithfulness. His promises make you unshakable, no matter the despair. #trustingod

Virtual AEHS Study 2:06:18

Virtual AEHS Study Day on 'Naming the Unspeakable: Faithfulness for a Planet in Crisis'

The Book of 0:07:15

The Book of Job's Wisdom on How God Runs the World

Комментарии
Автор

Congrats Doctor!! :-) Looking forward for your future work!

MikeAirforce
Автор

Thanks for sharing your work! Always great so see what you're up to!

DerPylz
Автор

Cool 😎 your explanation was very understandable

serta
Автор

Congrats Dr. Letitia!!!! Wow, YOU :-D :-) P.S. We missed you!!

beatrixcarroll
Автор

This sounds very useful! LLM users tend to assume that just because it writes like a human, that it can introspect and reason about its thought processes, which of course not a given. But it’s great to see progress on measuring this ability (or at least self-consistency) so that newer models can be more ergonomic.

alexkubiesa
Автор

Yay, new video!
Thanks for letting me pass yesterday lol

fingerstyledojo
Автор

Congrats on the PhD! This is really valuable work! I'm currently trying to squeeze out as much reasoning capabilities as I can out of small LLMs (7-15B) for my company's product, and I'd love a longer video or recorded talk going into details of your findings, any patterns you've found that contribute to improving or reducing self-consistency, or any insights on which existing models or training corpora result in better self consistency and reasoning capabilities. If you have any pointers, I'd appreciate it!

MaxShawabkeh
Автор

1 minute ago for non members ... good to see ya

nitinss
Автор

So you came up with a method, didn't have time to explain the method to us, and didn't show us that it works. Great.

If you still have time before Bangkok I would suggest rerecording and focusing on the implementation and interpretation of results rather than the context and wordy descriptions.

anluifb