arXiv:2608. 03222v1 Announce Type: cross Abstract: Software engineering (SWE) agents resolve repository-level issues through long trajectories that grow increasingly expensive as context accumulates.
Software engineering (SWE) agents resolve repository-level issues through long trajectories that grow increasingly expensive as context accumulates. Failed runs tend to be longer and exhibit redundant exploration or looping, suggesting that some failures may be detectable before completion.
TACIT-Switch is a cost‑aware routing method that decides when to hand off from a cheaper, smaller language‑model agent to a more expensive, larger one. It learns permanent handoff policies using Teacher‑Annotated Censored Intervention Times (TACIT), treating each annotation as an interval‑censored observation on a cumulative‑risk scale. In controlled simulations, TACIT‑Switch improves success rates by 7.4–11.1 percentage points over other routing baselines while keeping cost comparable, and it achieves the highest held‑out success on ALFWorld and DABench datasets.
arXiv:2606. 29654v1 Announce Type: new Abstract: Multi-agent deliberation among LLMs can improve reasoning, but deployment requires deciding when the current answer is reliable enough to act on and when it should be escalated to human review.
arXiv:2607. 19338v1 Announce Type: new Abstract: Coding agents increasingly operate in executable environments where a failed attempt produces actionable feedback rather than merely an incorrect answer.
The paper introduces a self‑healing harness that enforces admission control over language‑model agents’ self‑modifications. The harness runs a Detect‑Notice‑Heal‑Validate loop, allowing agents to propose rule changes that are only granted persistent authority after demonstrating improvement on a failure case without regressing on protected cases. Across 16 benchmark runs, the harness rejected many locally beneficial proposals that caused collateral regressions, while improving task‑completion scores and reliability.
ObserverBench is a benchmark framework that evaluates whether internal mechanistic estimators—called observers—are suitable for guiding interventions, control, or safety actions in language models. It separates estimation accuracy from the loss incurred by the chosen action, showing that accurate predictions do not always lead to better decisions. Experiments on GPT‑2‑small, Qwen2.5‑7B, Gemma‑2‑9B‑it, and Qwen3.5‑9B demonstrate that observers trained on action loss can reduce deployment loss, while traditional metrics like AUROC may rank monitors differently from actual performance.
Large language model agents depend on external harnesses to exchange information with their environment and to recover from execution errors, but recovery is typically evaluated only by overall task success, masking a key trade‑off. The authors treat recovery as a causal decision problem, comparing outcomes with and without recovery from the same execution state to separate rescue from harm and analyze how its value evolves over time. They propose the Causal Intervention Router (CIR), a lightweight policy that uses pre‑recovery information to decide when intervention is beneficial, achieving a 3‑point increase in success on long‑horizon ALFWorld tasks with Qwen3‑14B while preserving correct observations and demonstrating that recovery’s benefit is not solely due to new observations.
arXiv:2607. 06503v1 Announce Type: new Abstract: Large language model (LLM) agents solving multi-step tasks frequently commit to trajectories that are doomed to fail, yet continue to consume substantial inference compute before the failure becomes observable.
Coding agents increasingly operate in executable environments where a failed attempt produces actionable feedback rather than merely an incorrect answer. Existing cost-aware systems typically treat such failures as cascade decisions: try a cheap model first, then escalate hard cases to a stronger and more expensive model.
ObserverBench is a benchmark framework that evaluates whether internal mechanistic estimators—called observers—are suitable for guiding interventions, control, or safety actions in language models. It separates estimation accuracy from the loss incurred by the chosen action, demonstrating that accurate average estimates can still lead to poor decisions. Experiments on GPT‑2‑small, Qwen2.5‑7B, Gemma‑2‑9B‑it, and Qwen3.5‑9B show that observers trained on action loss tend to select lower‑loss actions, while traditional metrics like AUROC can rank monitors differently from deployment loss, highlighting the need for task‑specific evaluation.
arXiv:2609.36138v1 Announce Type: new Abstract: Before invoking external tools, an agentic LLM must select among a K-way action space: executing a call, seeking clarification, answering directly, or...