specific method · filed under post-training
Token-budget-based SFT data blending
Constructing an SFT data mixture using target proportions of tokens rather than counts of examples.
Also called token-level target proportions.
- source
- 1
- model
- 1
- 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.
We construct the final SFT mixture using token-level target proportions rather than example counts
usedunclearin Nemotron 3 UltraNVIDIA
Filed alongside
Other methods under post-training :: supervised fine-tuning.
Supervised fine-tuningFine-tuningLight SFT on teacher-distribution dataLoRA fine-tuningRejection samplingSelf-correction cold startSFT bootstrapping with synthetic dataAdversarial fine-tuningAutoregressive fine-tuningCold-started SFT modelEvaluation-based early stoppingFine-tuning with demographically balanced datasetsFrozen-network warmup for new special tokensInstruction hierarchy trainingInstruction-following fine-tuningLightweight speaker fine-tuningPrompt-based cold start for tool useReasoning-data system promptSafety training to an internal specificationSFT checkpoint for RL researchSFT on MiMo-generated task dataSpecialist-model trainingSupervised fine-tuning at 256k sequence length