pith. sign in

This incurs additional training cost: for the 0.5B model, AR SFT takes 15 H200-hours while MARS takes 33 H200-hours (2.2×); for the 7B model, 100 vs 202 H200-hours (2.0×)

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

fields

cs.CL 1

years

2026 1

verdicts

UNVERDICTED 1

representative citing papers

MARS: Enabling Autoregressive Models Multi-Token Generation

cs.CL · 2026-04-08 · unverdicted · novelty 7.0

MARS fine-tunes autoregressive models to predict multiple tokens per step via continued training on instruction data, achieving 1.5-1.7x throughput while matching baseline accuracy and supporting real-time speed adjustment.

citing papers explorer

Showing 1 of 1 citing paper.

  • MARS: Enabling Autoregressive Models Multi-Token Generation cs.CL · 2026-04-08 · unverdicted · none · ref 18

    MARS fine-tunes autoregressive models to predict multiple tokens per step via continued training on instruction data, achieving 1.5-1.7x throughput while matching baseline accuracy and supporting real-time speed adjustment.