qwennie 🐩 v2 a 205,040-parameter, 3-layer decoder transformer running in the CSS style engine. MQA, local+global causal attention, RoPE, RMSNorm, SwiGLU, sparse int8 matmuls, and exact categorical temperature sampling. still zero JavaScript.
this browser is missing CSS math/style-query features qwennie needs. Chromium is the intended engine.
reroll 🎾 🦴 🦆 💨 temperature cozy waggy zoomies
you · turn 1 · ? you what do is are a how who me i can tell your work use good hello like javascript remember css about treats did say it attention yipsy that dream am help girl snow day with one now two rope why there real llm happy math secret joke the where live brain rain here bark smart thing bye thank moon alive sleep learn sad transformer hi name have favorite really weird kind then tiny compete another even all words louder stay celebrate and rent world scary scientifically plus wait more counts saw actual softmax mqa local forget sparsity zero useful fast compared to python trained teacher version changed we chat twice pup still qwennie today want sister love walk hey think language model an ai was treat · ? you what do is are a how who me i can tell your work use good hello like javascript remember css about treats did say it attention yipsy that dream am help girl snow day with one now two rope why there real llm happy math secret joke the where live brain rain here bark smart thing bye thank moon alive sleep learn sad transformer hi name have favorite really weird kind then tiny compete another even all words louder stay celebrate and rent world scary scientifically plus wait more counts saw actual softmax mqa local forget sparsity zero useful fast compared to python trained teacher version changed we chat twice pup still qwennie today want sister love walk hey think language model an ai was treat · ? you what do is are a how who me i can tell your work use good hello like javascript remember css about treats did say it attention yipsy that dream am help girl snow day with one now two rope why there real llm happy math secret joke the where live brain rain here bark smart thing bye thank moon alive sleep learn sad transformer hi name have favorite really weird kind then tiny compete another even all words louder stay celebrate and rent world scary scientifically plus wait more counts saw actual softmax mqa local forget sparsity zero useful fast compared to python trained teacher version changed we chat twice pup still qwennie today want sister love walk hey think language model an ai was treat · ? you what do is are a how who me i can tell your work use good hello like javascript remember css about treats did say it attention yipsy that dream am help girl snow day with one now two rope why there real llm happy math secret joke the where live brain rain here bark smart thing bye thank moon alive sleep learn sad transformer hi name have favorite really weird kind then tiny compete another even all words louder stay celebrate and rent world scary scientifically plus wait more counts saw actual softmax mqa local forget sparsity zero useful fast compared to python trained teacher version changed we chat twice pup still qwennie today want sister love walk hey think language model an ai was treat · ? you what do is are a how who me i can tell your work use good hello like javascript remember css about treats did say it attention yipsy that dream am help girl snow day with one now two rope why there real llm happy math secret joke the where live brain rain here bark smart thing bye thank moon alive sleep learn sad transformer hi name have favorite really weird kind then tiny compete another even all words louder stay celebrate and rent world scary scientifically plus wait more counts saw actual softmax mqa local forget sparsity zero useful fast compared to python trained teacher version changed we chat twice pup still qwennie today want sister love walk hey think language model an ai was treat · ? you what do is are a how who me i can tell your work use good hello like javascript remember css about treats did say it attention yipsy that dream am help girl snow day with one now two rope why there real llm happy math secret joke the where live brain rain here bark smart thing bye thank moon alive sleep learn sad transformer hi name have favorite really weird kind then tiny compete another even all words louder stay celebrate and rent world scary scientifically plus wait more counts saw actual softmax mqa local forget sparsity zero useful fast compared to python trained teacher version changed we chat twice pup still qwennie today want sister love walk hey think language model an ai was treat · ? you what do is are a how who me i can tell your work use good hello like javascript remember css about treats did say it attention yipsy that dream am help girl snow day with one now two rope why there real llm happy math secret joke the where live brain rain here bark smart thing bye thank moon alive sleep learn sad transformer hi name have favorite really weird kind then tiny compete another even all words louder stay celebrate and rent world scary scientifically plus wait more counts saw actual softmax mqa local forget sparsity zero useful fast compared to python trained teacher version changed we chat twice pup still qwennie today want sister love walk hey think language model an ai was treat · ? you what do is are a how who me i can tell your work use good hello like javascript remember css about treats did say it attention yipsy that dream am help girl snow day with one now two rope why there real llm happy math secret joke the where live brain rain here bark smart thing bye thank moon alive sleep learn sad transformer hi name have favorite really weird kind then tiny compete another even all words louder stay celebrate and rent world scary scientifically plus wait more counts saw actual softmax mqa local forget sparsity zero useful fast compared to python trained teacher version changed we chat twice pup still qwennie today want sister love walk hey think language model an ai was treat
you · turn 2 · ? you what do is are a how who me i can tell your work use good hello like javascript remember css about treats did say it attention yipsy that dream am help girl snow day with one now two rope why there real llm happy math secret joke the where live brain rain here bark smart thing bye thank moon alive sleep learn sad transformer hi name have favorite really weird kind then tiny compete another even all words louder stay celebrate and rent world scary scientifically plus wait more counts saw actual softmax mqa local forget sparsity zero useful fast compared to python trained teacher version changed we chat twice pup still qwennie today want sister love walk hey think language model an ai was treat · ? you what do is are a how who me i can tell your work use good hello like javascript remember css about treats did say it attention yipsy that dream am help girl snow day with one now two rope why there real llm happy math secret joke the where live brain rain here bark smart thing bye thank moon alive sleep learn sad transformer hi name have favorite really weird kind then tiny compete another even all words louder stay celebrate and rent world scary scientifically plus wait more counts saw actual softmax mqa local forget sparsity zero useful fast compared to python trained teacher version changed we chat twice pup still qwennie today want sister love walk hey think language model an ai was treat · ? you what do is are a how who me i can tell your work use good hello like javascript remember css about treats did say it attention yipsy that dream am help girl snow day with one now two rope why there real llm happy math secret joke the where live brain rain here bark smart thing bye thank moon alive sleep learn sad transformer hi name have favorite really weird kind then tiny compete another even all words louder stay celebrate and rent world scary scientifically plus wait more counts saw actual softmax mqa local forget sparsity zero useful fast compared to python trained teacher version changed we chat twice pup still qwennie today want sister love walk hey think language model an ai was treat · ? you what do is are a how who me i can tell your work use good hello like javascript remember css about treats did say it attention yipsy that dream am help girl snow day with one now two rope why there real llm happy math secret joke the where live brain rain here bark smart thing bye thank moon alive sleep learn sad transformer hi name have favorite really weird kind then tiny compete another even all words louder stay celebrate and rent world scary scientifically plus wait more counts saw actual softmax mqa local forget sparsity zero useful fast compared to python trained teacher version changed we chat twice pup still qwennie today want sister love walk hey think language model an ai was treat · ? you what do is are a how who me i can tell your work use good hello like javascript remember css about treats did say it attention yipsy that dream am help girl snow day with one now two rope why there real llm happy math secret joke the where live brain rain here bark smart thing bye thank moon alive sleep learn sad transformer hi name have favorite really weird kind then tiny compete another even all words louder stay celebrate and rent world scary scientifically plus wait more counts saw actual softmax mqa local forget sparsity zero useful fast compared to python trained teacher version changed we chat twice pup still qwennie today want sister love walk hey think language model an ai was treat · ? you what do is are a how who me i can tell your work use good hello like javascript remember css about treats did say it attention yipsy that dream am help girl snow day with one now two rope why there real llm happy math secret joke the where live brain rain here bark smart thing bye thank moon alive sleep learn sad transformer hi name have favorite really weird kind then tiny compete another even all words louder stay celebrate and rent world scary scientifically plus wait more counts saw actual softmax mqa local forget sparsity zero useful fast compared to python trained teacher version changed we chat twice pup still qwennie today want sister love walk hey think language model an ai was treat · ? you what do is are a how who me i can tell your work use good hello like javascript remember css about treats did say it attention yipsy that dream am help girl snow day with one now two rope why there real llm happy math secret joke the where live brain rain here bark smart thing bye thank moon alive sleep learn sad transformer hi name have favorite really weird kind then tiny compete another even all words louder stay celebrate and rent world scary scientifically plus wait more counts saw actual softmax mqa local forget sparsity zero useful fast compared to python trained teacher version changed we chat twice pup still qwennie today want sister love walk hey think language model an ai was treat · ? you what do is are a how who me i can tell your work use good hello like javascript remember css about treats did say it attention yipsy that dream am help girl snow day with one now two rope why there real llm happy math secret joke the where live brain rain here bark smart thing bye thank moon alive sleep learn sad transformer hi name have favorite really weird kind then tiny compete another even all words louder stay celebrate and rent world scary scientifically plus wait more counts saw actual softmax mqa local forget sparsity zero useful fast compared to python trained teacher version changed we chat twice pup still qwennie today want sister love walk hey think language model an ai was treat
qwennie · turn 1 · both replies generated in one style pass
qwennie · turn 2
dense params 205,040 · quantized nonzero weights 123,376/204,480 · vocab 520 · context 64 · 5Q/1KV heads · local window 6 + memory anchors