#Fine-Tuning
does anyone actually do fine tuning on LLMs anymore. I mean beyond experiments/fun
November 16, 2026 at 11:54 PM
Fine-tuning LLMs is the new sourdough starter
November 18, 2026 at 2:30 PM
oh my gourd (i am fine tuning this later)
November 18, 2026 at 4:51 AM
Horizon Forbidden West @guerrilla.bsky.social

Still fine-tuning the whole tool thanks to the feedback I received.

Say tuned!

Nota bene: I know I said something years ago :-)

#horizonforbiddenwest #horizon #aloy #GuerrillaGames #virtualphotography #vpsupport #vpinspired
November 17, 2026 at 7:24 AM
Are you fine-tuning LLMs?
November 19, 2026 at 2:43 AM
There computers equipment still to be connected and I’m sure there will be fine tuning needed. But I have a workspace!!

I think aside from a celebratory glass of prosecco and a shower I am done for the day.
November 15, 2026 at 9:00 PM
A legislative leader is all about inside baseball stuff: uniting a caucus, fine-tuning legislation for conflicting internal demands, seating members fairly in committees, coordinating and prioritizing investigations, working with outside groups, raising money, raising money, and also raising money.
November 18, 2026 at 4:35 PM
Theoretical physicist Prof Dan Hooper explains why fine tuning is a sensible - but imperfect - signpost which physicists can use to focus theoretical efforts.

Would you accept your friend winning the lottery 3 times or ask if you'd missed something?

Link: www.youtube.com/watch?v=sesx...

⚛️🧪
The Attack of the Science Populists: Naturalness And Fine Tuning Are Signposts
YouTube video by Bad Boy of Science
www.youtube.com
November 16, 2026 at 1:39 PM
We totally have to crank up the development pace to compete with the onslaught of plugins made with AI... but we are doing it completely by hand, carefully designing every feature and fine-tuning every single formula. However... does anyone care anymore? We surely hope so.
November 18, 2026 at 5:06 PM
Researchers present Ambient SFT, a new framework that improves large language models' fine-tuning by using high-temperature predictions from a frozen pre-trained model, enhancing performance in data-scarce scenarios. https://arxiv.org/abs/2611.23229
Ambient LLMs: Data-Efficient Supervised Fine-tuning via Temperature Matching
ArXiv link for Ambient LLMs: Data-Efficient Supervised Fine-tuning via Temperature Matching
arxiv.org
November 19, 2026 at 5:10 AM
I think we’re seeing Skynet / Matrix doing rehearsals and fine-tuning.
November 17, 2026 at 4:32 PM
Fine-tuning the AI for these scary hornet bots...definitely scarier when I plop down a few of them at once!

I'm tempted to use some boid logic to space them out, but instead I'm just randomly moving their desired position every few seconds (except when attacking).

#godot #gamedev #indiedev
November 17, 2026 at 10:41 PM
Plant biotechnologists are fine-tuning genetic pathways to develop a new generation of crops equipped to cope with the myriad pressures of climate change www.nature.com/articles/s41...
#COP31 [🔓read free with login]
The seeds of resilience - Nature Biotechnology
Plant biotechnologists are fine-tuning genetic pathways to develop a new generation of crops equipped to cope with the myriad pressures of climate change.
www.nature.com
November 18, 2026 at 12:36 AM
This is a small drawing of Equinox as a teenager. I was primarily focused on locking down his silhouette and fine-tuning small elements of his design.

#art #illustration #characterdesign
November 17, 2026 at 7:08 AM
Would they not be doing fine tuning with synthID active?
November 14, 2026 at 5:24 PM
Fine-Tuning an Order-Flow Foundation Model for Sub-Second Forecasting arxiv.org/abs/2611.14117 #QuantLinkADay
November 19, 2026 at 9:31 AM
one problem with making an Outer Wilds style game (does that have a genre name yet? i refuse to acknowledge the term "metroidbrainia") is the puzzles require a lot of fine-tuning with playtesters... but each person can only playtest each puzzle once
November 16, 2026 at 8:30 PM
The #FinancialSector - a threat to the security of us all.

“All they’ve learnt is [that their] risk management needs a little more fine tuning”.

www.ft.com/content/1fcb...
‘Situational Awareness 2.0’: Leopold Aschenbrenner plots comeback
The 25-year-old AI trader has held talks about employing leverage again and raising funds after wipeout nearly toppled firm
www.ft.com
November 19, 2026 at 12:16 AM
The difference bw a 7b base model and Claude opus isn’t just fine-tuning through.
November 14, 2026 at 6:21 PM
LoRA fine-tuning on a single GPU: a step-by-step guide

LoRA fine-tuning on a single GPU lets you adapt a small language model (SLM — a model from a few million up to 10 billion parameters that runs on consumer hardware) to your own task…
LoRA fine-tuning on a single GPU: a step-by-step guide
Learn to fine-tune LLMs with LoRA on a single consumer GPU: dataset, hyperparameters, and launching training.
ai-release.org
November 15, 2026 at 10:06 AM
Was fine tuning my physics on Kiki a little - have a little preview <3

Inspiration for this movement from a couple CC's - I love soft booodiieees
November 11, 2026 at 6:24 PM
💡 Questions for the Community: In 2026, when is fine-tuning actually worth the maintenance overhead over dynamic few-shot prompting?

Have you seen genuine ROI from fine-tuned weights for agents?

#ai #machinelearning #llm #data
November 19, 2026 at 5:58 AM
I don’t do meanings behind names. I find what feels close and then fine tune it until the character starts opening up in my mind. Sometimes it feels a bit like tuning into a radio station.

#WriteSky #WritingCommunity
#WritingPrompts
Tell me about your #AwesomeCharacters

Technical question today!

Do your character names have meanings behind them? Etymologically or otherwise?

Or is it more about how the names feel to you?

Have you ever looked up the meaning after the fact?

#WriteSky #WritingCommunity #WritingPrompts
November 18, 2026 at 3:37 PM
yeah, too bad good policy ain't easy ... they should have had the guts to make it work rather than cut and run.

evidence-based policy is supposed to be about iterating and fine tuning
November 17, 2026 at 11:43 PM