transformer · 72 topics

elastic threshold attention long context decoding

modality discrepancy transformer for ambivalence and hesitancy recognition

subliminal prompting causal depth multi token confounds

dantinox a unified framework for multi paradigm language modeling

latent undertow how ordinary typos break probes

ncp archpreview latent space language models

auk technical report

transformer logical consistency qk alignment

uno lossless speedups via discrete diffusion

gated deltanet nvfp4 w4a4 minima

latentpress context compression

smelt scaling laws for compute matched moe looped transformers

random attention rethinking kv cache eviction for efficient reasoning

ai weather model precipitation forecasts imerg

neomme single tower multimodal encoder

the mask is not the model

luce relightable gaussians for 3d asset generation

eduriskx neuro symbolic framework

large models for battery prognostics and health management

prefix sliding

how much rank does lora need rank error bounds for transformer attention

triplu trilinear product ffns tiny language models

transformer models text summarization

fractional decay kv cache

from bert to frontier agents

depth aware sensitivity analysis of moe models

modular cognitive architecture emerges in large language models

uniswap streaming audio visual identity swapping

latent to 4d

thought aware kv cache compaction for reasoning via adaptive attention matching

dual flow transformers

llm cot serial depth bottleneck

training variable long sequences with data centric parallel

mass multiplayer world models

timestep conditioned transformers for global weather forecasting

memory decoder at scale

fusionml cpu gpu co execution apple silicon

sana video 2 0

a controlled study of attention only transformers

matrix recurrent units attention alternative

itnet unified integral transform

2605.19269

2605.15514

arxiv 2604 22771

autocompress critical layer isolation

20251006 linear regression icl

openarch pytorch llm architectures

termy terminal assistant

llm from scratch

microsoft retrochimera drug synthesis

type safe ai jev model

sakana ai frontier intelligence group

minicpm5 2b open sota

vericache lossless kv cache

simpledesign joint model for protein sequence and structure codesign

openai community gpt2

spark x25 4b release

neomme multimodal encoder

small transformer arc agi 1

how to build a diffusion language model

chatgpt logits softmax

google deepmind agi to asi

transformer transformer motion conditioned robot co design

discoformer

words are a byproduct of consciousness

which tokens does a hybrid model predict better

2026 06 07 papers transformers inherently succinct 2024

nvidia transaction foundation models

olmoearth v1 1 efficiency

transformer scalability crisis

llm architectures kv sharing mhc

etched valuation 10 3b