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.
Stories per day · 30 days
1 RSS
Relatedagentic-rl1ai-agents1benchmark1browsecomp1coding-agents1llm-agents1multi-session1pi-harness1prompt-injection1provenance1qwen1reinforcement-learning1rule-compliance1speculative-verification1subagents1
9 stories, newest first · one card per story
Stories tagged #context-compaction
Benchmark introduced for LLM agent rule compliance after context compaction
ARCBench measures violations in agent actions on the pi harness under different compaction strategies.
Context compaction in coding agents can launder forged user instructions
Flattening harnesses accept forged approvals 87-100% of the time, while four role-typed harnesses reject the same attack.
Mosaic: Multi-Session Agentic Reinforcement Learning
Mosaic improves multi-session agentic RL by crediting session contributions instead of broadcasting trajectory advantages.
REMORY: Learning Residual Memory for Context Compaction
REMORY adds residual soft memory tokens so agents compact context without losing later decision quality.
Does an Agent's History Tell You When Compaction Will Hurt? A Modest, Bounded Effect on the TRACE Paired-Replay Corpus
On TRACE's 590 AppWorld boundaries, recent agent history only weakly predicts harmful context compaction.
AutoCompact: Learning When to Compact Context in Long-Horizon Coding Agents
AutoCompact trains coding agents to compact context, raising SWE-bench Verified pass rates by 9.2 points.
KV-streams for Efficient Compaction in Agentic Reinforcement Learning
KV-streams stream the KV cache through compaction, speeding agentic RL training 2.6 to 5 times without hurting performance.
CliffCompaction: Cost-Efficient Compaction for Long-Horizon Coding Agents
CliffCompaction cuts long-horizon coding-agent context cost by up to 50% without rewriting retained text.