SpecCompact accelerates LLM agent context compaction with slot-parallel speculation
SpecCompact delivers 2.04-5.63x faster context compaction on AppWorld and τ³ benchmarks by parallel slot generation and target-model verification.
1 stories, newest first · one card per story
SpecCompact delivers 2.04-5.63x faster context compaction on AppWorld and τ³ benchmarks by parallel slot generation and target-model verification.
Today 1.6K readers · 2.5K views · 778 reads · all time 189K views