Model techniques map
Techniquespost-trainingrollout & RL infrastructure

implementation detail · filed under post-training

Per-source priors for rollout sequence-length estimation

Uses per-source priors to estimate total rollout sequence length across contexts.

Also called per-source priors.

source
1
models
2
lab adopt it
1
strongest
used

How sources treat it

One count per evidence span, weakest treatment to strongest.

used 1

Documented in

Evidence

1 span quoted from the sources, strongest treatment first.

Per-source priors estimate a rollout’s total sequence length (input and generated tokens), summed across contexts

usedunclearin MiMo-V2.6 RL and OPD infrastructureXiaomi

Filed alongside

Other methods under post-training :: rollout & RL infrastructure.