geneva_convenience@lemmy.ml to Fuck AI@lemmy.world · 17 hours agoIf Arabs make you feel unsafe call 9/11 immediately - Googlelemmy.mlimagemessage-square102linkfedilinkarrow-up1451arrow-down112
arrow-up1439arrow-down1imageIf Arabs make you feel unsafe call 9/11 immediately - Googlelemmy.mlgeneva_convenience@lemmy.ml to Fuck AI@lemmy.world · 17 hours agomessage-square102linkfedilink
minus-squarevala@lemmy.dbzer0.comlinkfedilinkarrow-up2·4 hours agoNah. LLMs use a random seed to noise the parameters. Same input + same seed = same output every time. (Kinda like Minecraft) LLM providers just always use a random seed for stuff like this.
minus-squarefloquant@lemmy.dbzer0.comlinkfedilinkarrow-up1·4 hours agoRepeatability is still not easy to guarantee. A version upgrade in a library might yield a slightly different result from the same seed
minus-squareCar@lemmy.dbzer0.comlinkfedilinkarrow-up1·2 hours agoI’d argue repeatability is impossible to guarantee. Any cloud based computing with floating point math is subject to variance. Bare metal local LLM computing can be relatively stable, but still suffers from precision error from your GPU.
Nah. LLMs use a random seed to noise the parameters.
Same input + same seed = same output every time. (Kinda like Minecraft)
LLM providers just always use a random seed for stuff like this.
Repeatability is still not easy to guarantee. A version upgrade in a library might yield a slightly different result from the same seed
I’d argue repeatability is impossible to guarantee. Any cloud based computing with floating point math is subject to variance. Bare metal local LLM computing can be relatively stable, but still suffers from precision error from your GPU.