KV-cache compression

OPERA

OPERA: Alleviating Hallucination in Multi-Modal Large Language Models via Over-Trust Penalty and Retrospection-Allocation

Superseded baseline#185 of 234 most-superseded · first seen Nov 29, 2023

Cited as a baseline — critiqued by newer work, not yet beaten on a benchmark here

1 papers critique it · 0 beat it on benchmarks

What papers say

Verbatim critique sentences, each from a paper that cites OPERA as a baseline.

However, all above mentioned works introduce additional computational overhead, thus slows down model inference speed.
PruneHal: Reducing Hallucinations in Multi-modal Large Language Models through Adaptive KV Cache Pruning

What to use instead

Recent methods in the same sub-problem, not yet superseded in the knowledge base — arXiv benchmark leaders, not vetted production recommendations.