gpjt/1xrtx3090-stacked-interventions
Text Generation • 0.2B • Updated • 313
Giles Thomas
gpjt
AI & ML interests
Doing my best to speedrun 20 years of AI research. YMMV
Recent Activity
updated a collection 5 days ago
LLM from scratch updated a model 5 days ago
gpjt/1xrtx3090-stacked-interventions published a model 5 days ago
gpjt/1xrtx3090-stacked-interventionsOrganizations
None yet
gpjt/1xrtx3090-baseline
Text Generation • 0.2B • Updated • 217
gpjt/8xa100m40-stacked-interventions-3
Text Generation • 0.2B • Updated • 209
gpjt/8xa100m40-stacked-interventions-2
Text Generation • 0.2B • Updated • 264
gpjt/8xa100m40-stacked-interventions-1
Text Generation • 0.2B • Updated • 255
gpjt/8xa100m40-baseline-8
Text Generation • 0.2B • Updated • 524
gpjt/8xa100m40-baseline-7
Text Generation • 0.2B • Updated • 523
gpjt/8xa100m40-baseline-6
Text Generation • 0.2B • Updated • 513
gpjt/8xa100m40-baseline-4
Text Generation • 0.2B • Updated • 325
gpjt/8xa100m40-baseline-5
Text Generation • 0.2B • Updated • 509
gpjt/8xa100m40-baseline-3
Text Generation • 0.2B • Updated • 580
gpjt/8xa100m40-baseline-2
Text Generation • 0.2B • Updated • 533
gpjt/8xa100m80-no-amp
Text Generation • 0.2B • Updated • 263
gpjt/8xa100m40-weight-decay-cerebras
Text Generation • 0.2B • Updated • 363
gpjt/8xa100m40-weight-decay-gpt2
Text Generation • 0.2B • Updated • 349
gpjt/8xa100m40-qkv-bias
Text Generation • 0.2B • Updated • 94
gpjt/8xa100m40-schedule-learning-rate
Text Generation • 0.2B • Updated • 94
gpjt/8xa100m40-remove-dropout
Text Generation • 0.2B • Updated • 98
gpjt/8xa100m40-baseline
Text Generation • 0.2B • Updated • 93
gpjt/1xrtx3090m24-fineweb-edu-2x
Text Generation • 0.2B • Updated • 88
gpjt/1xrtx3090m24-fineweb-edu
Text Generation • 0.2B • Updated • 89
gpjt/1xrtx3090m24-fineweb
Text Generation • 0.2B • Updated • 98
gpjt/8xh100m80-latest
Text Generation • 0.2B • Updated • 88
gpjt/8xh100m80-best
Text Generation • 0.2B • Updated • 94
gpjt/8xb200m160
Text Generation • 0.2B • Updated • 98
gpjt/8xa100m80
Text Generation • 0.2B • Updated • 96
gpjt/8xa100m40
Text Generation • 0.2B • Updated • 130
gpjt/8xa100m40-gradient-clipping
Text Generation • 0.2B • Updated • 90
gpjt/test5
Text Generation • 0.2B • Updated • 8
gpjt/test4
Text Generation • 0.2B • Updated • 6