Live data from Hacker News

CPUs Aren't Dead. Gemma2B Out Scored GPT-3.5 Turbo on Test That Made It Famous

seqpu.com

11–20 of 59 posts

Re: CPUs Aren't Dead. Gemma2B Out Scored GPT-3.5 Turbo on Test That Made It Famous

#11
post #9
post #7

> The model does not need to be retrained. It needs surgical guardrails at the exact moments where its output layer flinches. > With those guardrails — a calculator for arithmetic, a logic solver for formal puzzles, a per-requirement verifier for structural constraints, and a handful of regex post-passes — the projected score climbs to ~8.2. Surgical guardrails? Tools, those are just tools.

"Surgical "is the kind of wordage that LLMs seem to love to output. I have had to put in my .md file the explicit statement that the word "surgical" should only be used when referring to an actual operation at the block...

you're right, they are tools. that's kind of the point. PAL is a subprocess that runs a python expression. Z3 is a constraint solver. regex is regex. calling them "surgical" is just about when they fire, not what they are. the model generates correctly 90%+ of the time. the guardrails only trigger on the 7 specific patterns we found in the tape. to be clear, the ~8.0 score is the raw model with zero augmentation. no tools, no tricks. just the naive wrapper. the guardrail projections are documented separately. all the code is in the article for anyone who wants to review it.

Re: CPUs Aren't Dead. Gemma2B Out Scored GPT-3.5 Turbo on Test That Made It Famous

#13
post #6
post #5

Earlier quoted context omitted.

But GPT-3.5 was benchmaxxing too.

GPT 3.5 Turbo knowledge cutoff was circa 2021. MT-Bench is from 2023. Not suggesting improvements on small models aren't possible (or forthcoming, the 1.85 bit etc models look exciting), but this almost certainly isn't that.

[dead]

Re: CPUs Aren't Dead. Gemma2B Out Scored GPT-3.5 Turbo on Test That Made It Famous

#14
post #8

I yearn for the days when I can program on my PC with a programming llm running on the CPU locally.

you're honestly not that far off. the coding block on this model scored 8.44 with zero help. it caught a None-init TypeError on a code review question that most people would miss. one question asked for O(n) and it just went ahead and shipped O(log(min(m,n))) on its own. it's not copilot but it's free, it's offline, and it runs on whatever you have. there's a 30-line chat.py in the article you can copy and run tonight.

Re: CPUs Aren't Dead. Gemma2B Out Scored GPT-3.5 Turbo on Test That Made It Famous

#15

Posters comment is dead. It may be llm-assisted but should prob be vouched for anyway as long as the story isn't flagged.

appreciate the vouch but come on lol. we ran 80 questions, graded 160 turns by hand, documented 7 error classes, open sourced all the code, and put a live bot up for people to test. to write this post up took me hours. everyone is a critic lol.

Re: CPUs Aren't Dead. Gemma2B Out Scored GPT-3.5 Turbo on Test That Made It Famous

#18
post #9
post #7

> The model does not need to be retrained. It needs surgical guardrails at the exact moments where its output layer flinches. > With those guardrails — a calculator for arithmetic, a logic solver for formal puzzles, a per-requirement verifier for structural constraints, and a handful of regex post-passes — the projected score climbs to ~8.2. Surgical guardrails? Tools, those are just tools.

"Surgical "is the kind of wordage that LLMs seem to love to output. I have had to put in my .md file the explicit statement that the word "surgical" should only be used when referring to an actual operation at the block...

The core issue is that the LLM is using rhetoric to try to convince or persuade you. That's what you need to tell it not to do.

Re: CPUs Aren't Dead. Gemma2B Out Scored GPT-3.5 Turbo on Test That Made It Famous

#19
post #7

> The model does not need to be retrained. It needs surgical guardrails at the exact moments where its output layer flinches. > With those guardrails — a calculator for arithmetic, a logic solver for formal puzzles, a per-requirement verifier for structural constraints, and a handful of regex post-passes — the projected score climbs to ~8.2. Surgical guardrails? Tools, those are just tools.

>It needs surgical guardrails at the exact moments where its output layer flinches.

This article is very clearly shitty LLM output. Abstract noun and verb combos are the tipoff.

It's actually quite horrible, it repeats lines from paragraph to paragraph.

Post reply on HN