SubiculumCode

In human memory, we often use a sense of familiarity to guide our memory decision, in the absence of explicit recollection of details. There is a whole memory literature about "recollection and familiarity" that dissociates the two cognitive processes, recollection which involves retrieval of specific details of an experience, and familiarity, which is a sense of memory strength, but absent of any qualitative detail. Familiarity is a faster process, and can often spur subsequent retrieval attempts that can lead to actual recollection..e.g. you see someone that seems familiar, but can't place where...and after a few moments, you remember who they were and where you had met them.

When measuring these processes, one approach has been to ask participants to provide confidence ratings. Recollection tends to lead to threshold-like, very high confident responses. Familiarity is more graded and continuous. Many then use a dual-process ROC model to identify the recollection and familiarity components, on average, of a person's memory of a memory test (see work by Andy Yonelinas).

This kind of work goes beyond memory, but applied to the general problem of how people judge their confidence in answers.

Its likely been applied to LLMs. A familiarity signal would probably be pretty easy to generate... The recollection kind of component might take some of those introspection type of approaches. For example, these papers, which I have not read, [1] https://arxiv.org/html/2603.17839v1 [2]https://arxiv.org/abs/2603.09250 might be getting at these ideas.

This might be relevant

very strange. for me, on Firefox Android, it was light gray text on white background....

I think one of the impetuses to not using the Old School Essentials rules was the Hasbro DnD debacle where they started making moves to undo the Open Gaming License (OGL) in 2023. The resulting Dolmenwood system cleanly removed the last vestiges of anything that might be subject to the OGL...e.g. certain spell names, monster names, etc.

I'm just working up the courage (and the people) to run it. I might have the face of a grognard, but I've only run a handful of sessions as a gamemaster / dungeonmaster. Between relative inexperience and the demands of my career in autism research career, I have not yet taken the plunge.

On the plus side for prep, the Dolmenwood books and pdfs are wonderfully organized...bringing the same expertise that Gavin Norman brought to Old School Essentials that had won him (and his team) wide acclaim.

Bare bones is a plus for me honestly. I'm fairly tired of crunchy mechanics that takes days and slow player interaction.

I am sure those systems deserve a modern renovation, but as a general matter, I think the particulars of game systems matter less, overall, when it is not a mechanics heavy system like Pathfinder and DnD 5e...or Rolemaster. It mostly comes down to rulings over rules...which is why rules lite systems, e.g. Mausritter, Knave, Into the Odd, have play in the OSR/NSR scene.

I do not know where on the crunch spectrum the systems you mentioned fall,

For a video with a fairly comprehensive overview, worth checking out:

https://youtu.be/hL_YAarvB5U?is=SQGjmmBP0ignvSy4

I was a backer and the books are lovely, and the setting is awesome.

Also this is my favorite t-shirt: https://necroticgnome.com/products/dolmenwood-goatlord-t-shi...

Also check out 3d6 Down the Line, an actual play worth listening to. https://youtu.be/xNWg_IcbHl8?is=54oqh6TbY5g6DP-B

It is not naive. Naivety is being overly reductionistic.

Who can read that text color and background combo? I had to turn on reader mode in Firefox.

Gymnastics? The argument was straightforward and rational.

Anthropic/Dario getting a lot of hate in this thread, but the argument he makes is rational if you believe for one second that AI is important to national security and that it can, or will soon be incredibly dangerous is misused. People pointing out that this position benefits his company might be overly reductionist in how they think the world works. I believe Dario and Anthropic about their belief that AI safety is paramount because they put serious dollars behind research to improve safety/alignment, not just lip service for others. Those resources could get spent elsewhere but do not. Certainly it doesn't at OpenAI to that degree.

I believe forced exercise actually is counterproductive in terms of the facts that exercise usually has on depression, at least in animal models that have looked at the differences in forced versus voluntary exercise. There's a whole literature out there I believe.

The content was 'you need to get your head checked'. That isn't tone, that directly implying that if you hold that position, there is something wrong with you. It's rude and unnecessary.

How about, is there reasonable suspicion that a crime has occurred?

But are not investigations typically under some reasonable suspicion that a crime had occurred?

Not a lawyer, but destruction of evidence would only be valid if there was first some reasonable suspicion of a crime? Is that right?

Does this mean California is more likely to get mega flooding?

Its not popular to work for. By popular, I mean people play it, at both the amateur and professional levels.

Funny. In writing, the advice is to "kill your darlings". Don't let a particular paragraph that you love a lot stop the whole story because it doesnt fit.

Welcome to the world of Chess a couple of decades ago...but look at chess now: more popular than ever.

I hope that this doesn't mean that you need write proofs while streaming Twitch. ;)

I want to replace my Lithium home battery, and what I really want to do it to move to one of the new sodium ion batteries..e.g like the one CATL is supposed to have at some point?

Anyone have a good idea when these will be available for consumers?

True. However, I think you are missing the larger point. Forget about coding for a bit. Think about biological research, for example. Being smart, being able to put a lot pieces together, is very important...but this will just help you make better hypotheses...you still need to do the slow and hard work of running physical experiments. Smartness does not lead to a singularity all on it's own

5 seems incredibly smart to me in my conversations today about some pretty niche ideas in.longitudinal modeling. It.felt.like a big step up.from 4.8, to me

Oh yeah? Who even talks about the Turing Test anymore? Half a decade ago,that was the informal benchmark.

It's already happened but no one wants to admit it

I do want to add, that I am pretty bummed if Opus 5 is going to refuse the tasks I have been using Opus 4.8 for (neuroimaging). Fable absolutely refuses anything close to toughing neuroscience.

same as it ever was. It seems your argument implies a belief that you should always use the best model. Others think that not all tasks require the absolute most powerful, expensive, model.

Because they are a private company and get to do what they want.

Well, probably not, but I am not engaging in socialization. I am using it to help me decide among competing statistical modeling approaches, or other aspects of my research, including coding, neuroimaging pipelines, and an assortment of other thorny issues that come with longitudinal/developmental neuroscience.