arXiv AI
Adapting Context Compression for Long-Horizon Agents with Counterfactual Continuations
Read the original on arXiv AI →The Flow has not summarised this story yet — read it at arXiv AI.
Related stories
arXiv Machine Learning
Aug 10
Toward Reliable Context Compression for Long-Horizon Agents: An Empirical Study of Execution Instability
arXiv:2608. 06503v1 Announce Type: new Abstract: Recurrent context compression controls context growth in long-horizon agents, but its behavioral effects remain poorly understood.
Hugging Face Trending Papers
Aug 2
Control Under Compression: Reliability Frontiers for Tool-Using Agents
Tool-using language-model agents are governed not only by task prompts but also by persistent system-side instructions that specify tools, arguments, policies, execution protocols, and recovery. Compressing these agent control contexts (ACCs) can reduce input cost and context use, yet existing prompt-compression evaluations do not reveal whether the resulting control remains operationally reliable.
arXiv AI
Aug 18
What Does Context Compression Cost an Agent? Interaction Costs Unrevealed by Task-Completion Metrics
arXiv:2608. 16370v1 Announce Type: new Abstract: Task completion is the standard metric for evaluating context compression, yet it is incomplete: compression can increase an agent's interaction cost by forcing it to reacquire dropped state while leaving completion statistically unchanged.
arXiv AI
Sep 24
StateComp: Learning When to Compress History in Long Horizon Agents
StateComp introduces a method for long‑horizon agents to decide when to compress historical interactions based on the current agent state, rather than relying on fixed windows or periodic schedules. The framework uses a two‑stage annotation process to create KEEP and READY labels, trains an imbalance‑aware router on frozen language model representations, and groups adjacent READY interactions into compact summaries. Experiments on WorkBuddyBench show that StateComp cuts agent and summarization tokens by 52.27% and speeds up representation extraction 12.67‑fold while preserving task performance.
arXiv AI
Aug 7
TRAJDEBUG: Tracing Error Lifecycle to Identify Critical Failures in Long-Horizon Agent Trajectories
arXiv:2608. 06346v1 Announce Type: new Abstract: LLM-based agentic systems have shown remarkable capabilities in complex domains, while suffering from cascading errors and difficulty in debugging.
arXiv Machine Learning
Jul 31
Benchmarking the Residual: What Long-Horizon Evaluations Add Beyond Matched Short-Task Performance
arXiv:2607. 27283v1 Announce Type: new Abstract: Long-horizon benchmarks often show that agents fail more as tasks become longer.