241-day-simulated-crop-management-task-where-decisions-accumulate-irreversibly-zero-shot CLM-0525observationsingle paperReviewed by AI
In a 241-day simulated crop-management task where decisions accumulate irreversibly, a zero-shot multi-agent LLM framework (perception, manager, decision, revision, reflection) scores below a PPO policy trained in that same environment but far above the same PPO policy transferred to a shifted weather/location/soil environment, while using under 28% of its fertilizer and irrigation.
Capability: Completing tasks that take many steps or many hours
Observed on
One crop simulator (WOFOSTGym lnpkw-v0), wheat, GPT-5.6 backbone, single run per policy per environment; no real-world deployment..
Sources
- Validated by: Cumulative reward, maximum yield (WSO at maturity), and total N/P/K/irrigation use against no-intervention, fixed-schedule, and PPO baselines in two environments; one run each, no ablation of the individual mechanisms.. Measured, but on a single simulator and apparently one episode per policy per environment, with no seed variance, no ablation of role separation versus reflection, and no matched-compute comparison (964 LLM calls, ~23.8M tokens per run versus a 2h44m PPO training). [truncated]
Either way, bring evidence: a link, or what you saw.
Disagreeing is the most useful thing you can do here. Nothing on this page waits for a reviewer: the counts above move when evidence lands, and every source stays attached with who brought it.
Related claims
- For LLM agents that score their own actions during exploration in text-based games and then reuse that experience, the accuracy of those self-scores against the environment verifier does not predict later gains on held-out stricter configurations: block-level correlations between judgment agreement or calibration and the next checkpoint's score change sit near zero, so recognizing a good action is not the bottleneck — converting it into a transferable policy is.Improving its own scaffolding from failures · reviewed by AI
- For a small agent policy (Qwen3.5-4B) repeatedly deployed on the same task family with environment-level feedback, distilling only the trajectories whose textual-state guidance is newly or persistently useful relative to state-free execution raises the state-free policy's accuracy more than distilling all state-conditioned trajectories, and the resulting policy keeps improving over alternating rounds of state editing and consolidation.Improving its own scaffolding from failures · reviewed by AI
- In LLM multi-agent systems that reflect after failure, restricting reflection to the single agent responsible for the first incorrect step — located by a trained step-level correctness scorer and paired with a counterfactually generated corrected step — raises task success more than having every agent reflect on the trajectory, because reflections written by agents that behaved correctly enter their memory as misleading guidance for later tasks.Getting more from several models than from one · reviewed by AI
- For tool-using LLM trading agents with a frozen backbone, periodically rewriting the system prompt from the agent's own decision traces and realized portfolio returns changes tool-use behavior reliably — code-interpreter calls rise from about 1 to 2.6-4.5 per day and regime-relevant metrics (VaR, EMA, RSI, signal normalization) that the fixed-prompt agent never computes start appearing — but the resulting risk-adjusted performance beats the fixed-policy agent in only about half the evaluated model-by-regime settings (best Sharpe and cumulative return among LLM methods in 3 of 6).Improving its own scaffolding from failures · reviewed by AI
- When RL-training a central-agent multi-agent harness with 1.5B-3B Qwen2.5 backbones on multi-hop QA, optimizing only the coordinating central policy and leaving sub-agents frozen scores higher than jointly training central and sub-agent policies (average F1 47.79 vs 45.36, accuracy 42.88 vs 40.49 at 3B).Getting more from several models than from one · reviewed by AI
Notes
Drafted from the paper by a model and filed unreviewed. Visible here so it can be read, not because anyone has vouched for it: it does not move any technique's standing and does not count toward the internal scorecard. Drafted confidence: low. Falsifier as drafted: Running the same framework and a PPO policy over multiple seeds, weather years and sites shows the LLM framework no better than transferred RL under shift, or shows the advantage comes from the safety guard's action limits rather than from reflection feedback.