Wire Benchmark Runner: PG19 LR PPL results auto-append to extrapolation curve and re-render plot dd2ea7e verified mlnomad commited on May 3
Add live Benchmark Runner tab: pick model + benchmark (wikitext, LR PPL, LAMBADA, ARC, HellaSwag, OBQA, PIQA), run on Spaces CPU 8ca1acd verified mlnomad commited on May 3
Add PG19 extrapolation curve tab (ctx 1k to 32k); only ctx=4096 populated, others pending eval 0748a23 verified mlnomad commited on May 3
Add goat-v-482m-c4-then-fineweb-seed0 (C4 to FineWeb continuation, eval pending) 8a8e7ee verified mlnomad commited on May 3
goat-v 482M C4 eval results: LR PPL 108 vs FineWeb 36.7 — long-range win was dataset, not architecture 63732fb verified mlnomad commited on May 1
Add goat-v-482m-c4-seed0 to model registry; reference benchmark row pending eval 804bec6 verified mlnomad commited on May 1
Add GOAT_VVV to model registry; new tabs: Kernel Scalars, Logit Lens, Pattern Arena, Long-Range Needle 9055503 verified mlnomad commited on Apr 28
Add GOAT_VVV to model registry; new tabs: Kernel Scalars, Logit Lens, Pattern Arena, Long-Range Needle 352c49e verified mlnomad commited on Apr 28
Add GOAT_VVV to model registry; new tabs: Kernel Scalars, Logit Lens, Pattern Arena, Long-Range Needle 8bcd023 verified mlnomad commited on Apr 27
Add GOAT_VVV to model registry; new tabs: Kernel Scalars, Logit Lens, Pattern Arena, Long-Range Needle 09cc531 verified mlnomad commited on Apr 27
Add GOAT_VVV to model registry; new tabs: Kernel Scalars, Logit Lens, Pattern Arena, Long-Range Needle cb004ef verified mlnomad commited on Apr 27
Add GOAT_VVV to model registry; new tabs: Kernel Scalars, Logit Lens, Pattern Arena, Long-Range Needle 6ea22d8 verified mlnomad commited on Apr 27
flaxchat demo: YatNMN vs GELU text generation + comparison 8c15380 verified mlnomad commited on Apr 17
mlnomad commited on 