Diverse, not Short: A Length-Controlled Self-Learning Framework for Improving Response Diversity of Language Models
Authors:
Vijeta Deshpande,
Debasmita Ghose,
John D. Patterson,
Roger Beaty,
Anna Rumshisky
Abstract:
Diverse language model responses are crucial for creative generation, open-ended tasks, and self-improvement training. We show that common diversity metrics, and even reward models used for preference optimization, systematically bias models toward shorter outputs, limiting expressiveness. To address this, we introduce Diverse, not Short (Diverse-NS), a length-controlled self-learning framework th…
▽ More
Diverse language model responses are crucial for creative generation, open-ended tasks, and self-improvement training. We show that common diversity metrics, and even reward models used for preference optimization, systematically bias models toward shorter outputs, limiting expressiveness. To address this, we introduce Diverse, not Short (Diverse-NS), a length-controlled self-learning framework that improves response diversity while maintaining length parity. By generating and filtering preference data that balances diversity, quality, and length, Diverse-NS enables effective training using only 3,000 preference pairs. Applied to LLaMA-3.1-8B and the Olmo-2 family, Diverse-NS substantially enhances lexical and semantic diversity. We show consistent improvement in diversity with minor reduction or gains in response quality on four creative generation tasks: Divergent Associations, Persona Generation, Alternate Uses, and Creative Writing. Surprisingly, experiments with the Olmo-2 model family (7B, and 13B) show that smaller models like Olmo-2-7B can serve as effective "diversity teachers" for larger models. By explicitly addressing length bias, our method efficiently pushes models toward more diverse and expressive outputs.
△ Less
Submitted 26 May, 2025; v1 submitted 22 May, 2025;
originally announced May 2025.
pi+- p differential cross sections at low energies
Authors:
H. Denz,
P. Amaudruz,
J. T. Brack,
J. Breitschopf,
P. Camerini,
J. L. Clark,
H. Clement,
L. Felawka,
E. Fragiacomo,
E. F. Gibson,
N. Grion,
G. J. Hofman,
B. Jamieson,
E. L. Mathie,
R. Meier,
G. Moloney,
D. Ottewell,
O. Patarakin,
J. D. Patterson,
M. M. Pavan,
S. Piano,
K. Raywood,
R. A. Ristinen,
R. Rui,
M. E. Sevior
, et al. (6 additional authors not shown)
Abstract:
Differential cross sections for pi- p and pi+ p elastic scattering were measured at five energies between 19.9 and 43.3 MeV. The use of the CHAOS magnetic spectrometer at TRIUMF, supplemented by a range telescope for muon background suppression, provided simultaneous coverage of a large part of the full angular range, thus allowing very precise relative cross section measurements. The absolute n…
▽ More
Differential cross sections for pi- p and pi+ p elastic scattering were measured at five energies between 19.9 and 43.3 MeV. The use of the CHAOS magnetic spectrometer at TRIUMF, supplemented by a range telescope for muon background suppression, provided simultaneous coverage of a large part of the full angular range, thus allowing very precise relative cross section measurements. The absolute normalisation was determined with a typical accuracy of 5 %. This was verified in a simultaneous measurement of muon proton elastic scattering. The measured cross sections show some deviations from phase shift analysis predictions, in particular at large angles and low energies. From the new data we determine the real part of the isospin forward scattering amplitude.
△ Less
Submitted 3 December, 2005;
originally announced December 2005.
