Ben Mayer (t11s)

@transmissions11

cs @mit, prev @sailresearchco @paradigm

2λ.com

Joined September 2020

1,065 Following

94.8K Followers

Replies Reposts Media

Pinned
Ben Mayer (t11s)
@transmissions11
Nov 3
Had the pleasure of helping with this! 🫡 I suspect there's historically been a big overhang in applied ZKP work due to the sheer implementation & verification complexity. But now thanks to LLMs, implementation is far easier and autoformalization is practical. Exciting times :)
Anthropic
@AnthropicAI
Nov 3
New Anthropic research: progress on provable inference for large language models. Here, we introduce FlashProve. Building on established cryptographic techniques, FlashProve allows developers to verify whether a given response came from a given AI model. It’s substantially more
28
Ben Mayer (t11s)
@transmissions11
Oct 29
Pretty sad how base model style "pass in some tokens and we'll predict the next ones" APIs are so hard to get a hold of now, let alone actual base model checkpoints. Using base models for turbocharged autocomplete is a really delightful experience, chat isn't always optimal UX!
8
Ben Mayer (t11s)
@transmissions11
Oct 23
Been learning about ZK-SNARKs lately! The core ideas behind modern hash-based schemes are surprisingly accessible & elegant 🧙‍♂️ Partly to teach myself, I wrote a post explaining them from scratch for people like me without a deep math background:
ZK-SNARKs from the Ground Up
From xn--2-kmb.com
23
Ben Mayer (t11s)
@transmissions11
Sep 14
I feel like it's still underdiscussed that seemingly all of the world's ~severe misalignment incidents as of now occurred in evaluation/RL? This is not what LW/MIRI/etc were expecting iiuc. Why aren't we seeing incidents in deployment? Is eval awareness that strong?
23
Ben Mayer (t11s)
@transmissions11
Sep 9
part 2/2 of my work this summer! TPU v6e has insane FLOPs/$ value, but performance out of the box on SGLang-JAX/vLLM-TPU is often lacking So we wrote our own collective matmul kernels in Pallas, pushing Gemma 4 31B prefill performance up to ~63% MFU(!) sailresearch.com/blog/tpu-v6e-g…
Sail Research
@sailresearchco
Sep 9
We nearly doubled MFU on TPU v6e serving Gemma 4 31B, taking prefill from 32% to 63% MFU, and 2x throughput. v6e has 2.5x less HBM and less than half the memory bandwidth of an H100. @transmissions11 wrote it up here: sailresearch.com/blog/tpu-v6e-g…
11

Ben Mayer (t11s)

Ben Mayer (t11s)’s posts