ENTITY Jelinek-Mercer

Jelinek-Mercer

PulseAugur coverage of Jelinek-Mercer — every cluster mentioning Jelinek-Mercer across labs, papers, and developer communities, ranked by signal.

Total · 30d

1

1 over 90d

Releases · 30d

0

0 over 90d

Papers · 30d

1

1 over 90d

TIER MIX · 90D

TOPICS

SENTIMENT · 30D

1 day(s) with sentiment data

RECENT · PAGE 1/1 · 1 TOTAL

TOOL · CL_129132 · Jul 7 · 04:00

Transformer models use Jelinek-Mercer and Dirichlet-style smoothing for in-context learning

Researchers have identified two complementary smoothing mechanisms within transformer models that are believed to underlie in-context learning. The first mechanism, observed at a finite attention-weight scale, acts as a…