Pith. sign in

Reinforcement learning from hu- man feedback research program,

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

1 Pith paper citing it

fields

cs.CL 1

years

2024 1

verdicts

REJECT 1

representative citing papers

Natural Language Fine-Tuning

cs.CL · 2024-12-29 · reject · novelty 4.0

NLFT weights each token by how much its probability shifts under different natural language prompts and claims to beat supervised fine-tuning with 50 examples, but the paper's loss equation and reported gains are internally inconsistent.

citing papers explorer

Showing 1 of 1 citing paper.

  • Natural Language Fine-Tuning cs.CL · 2024-12-29 · reject · none · ref 8

    NLFT weights each token by how much its probability shifts under different natural language prompts and claims to beat supervised fine-tuning with 50 examples, but the paper's loss equation and reported gains are internally inconsistent.