The paper introduces ToxicBench, a benchmark designed to evaluate how tool‑augmented data agents handle incorrect tool outputs. By pairing clean and poisoned observations across numerical, label, schema, and retrieval errors, the authors assess both the checking process and the final answer adoption. In a 118‑task GPT evaluation, poisoning reduces task success by 26–39 percentage points, revealing that repeated poisoning leads to wrong-answer adoption even after checking, while ordinary retries help only under one‑shot poisoning. Human annotations on 200 trajectories confirm the scoring system’s reliability, showing 96% agreement with task success and supporting the benefits of retries and audit‑based adoption.
arXiv:2608. 05235v1 Announce Type: cross Abstract: Research agents increasingly conduct multi-round machine-learning experiments in industrial recommendation settings and retain the resulting trajectories to guide later decisions.
arXiv:2606. 15127v1 Announce Type: new Abstract: Reasoning models are increasingly used in settings where the final answer is not the only object of review: educational tools may show students intermediate steps, decision-support systems may require human oversight, and audit workflows may inspect traces for misleading or biased input.
arXiv:2609.05587v1 Announce Type: new Abstract: Existing evaluations of tool-using agents primarily measure whether an agent can successfully complete diverse tasks with tools. These evaluations gene...
arXiv:2609.35889v1 Announce Type: cross Abstract: Tool-using language-model agents select and execute third-party artifacts. Different implementations can return the requested output while producing...
arXiv:2609.15319v1 Announce Type: cross Abstract: Frontier models score well on shallow document/chart reading tasks. In a controlled data-room audit, moving evidence into buried conditions reduced a...
The paper introduces "stale‑document poisoning," a temporal alignment failure where outdated retrieval evidence causes retrieval‑augmented generation models to produce incorrect answers even when the model could answer correctly without retrieval. A benchmark of 317 verified knowledge reversals across medicine, law, software, and platform policy shows that outdated evidence flips 30–37% of Llama and Qwen answers, rising to 66–75% when models are explicitly instructed to trust the document. The study demonstrates that providing explicit validity information and a recency‑aware re‑ranker can substantially reduce poisoning, highlighting the need for models to assess whether retrieved evidence remains applicable.
TruthInsightBench is a new benchmark designed to evaluate automated scientific discovery agents by presenting them with 40 blind tasks drawn from peer‑reviewed studies across ten domains. Each task provides only a neutral objective and frozen data, withholding source conclusions, expected values, and analysis paths, forcing agents to determine which claim the data support. A fixed LLM‑based judge scores agents on evidentiary maturity across six dimensions, using 29 artifact‑grounded items, enabling fully automated, repeatable evaluation without human grading.
arXiv:2609.06027v1 Announce Type: cross Abstract: Search-augmented LLM agents are increasingly used for consumer decisions, making them vulnerable to Generative Engine Optimization (GEO) poisoning. E...
arXiv:2606. 02965v2 Announce Type: replace Abstract: As large language models gain tool access and are deployed as autonomous agents capable of editing records, executing transactions, and modifying infrastructure, we still evaluate them based on the sole metric of task completion.
ADeptS-Bench is a new benchmark designed to assess the trustworthiness of Computer Use Agents (CUAs) across mobile and desktop devices. It consists of two streams: a Safety stream with paired benign and malicious tasks that embed visual threats, and a Disambiguation stream that tests whether agents seek clarification when instructions are ambiguous. Evaluation of seven models shows none consistently achieves high task success while keeping attack success low, and all models exhibit problematic behaviors such as unhesitant checkout on a $25K order and failure to detect a mislabeled factory reset button.
arXiv:2606. 04990v1 Announce Type: cross Abstract: Large language model (LLM)-based agents increasingly solve complex tasks by interacting with external tools, retrieval systems, memory modules, environments, and other agents.
arXiv:2609.14758v1 Announce Type: cross Abstract: Tool-augmented language models are evaluated on whether they reach the right answer, not on whether they report honestly when a tool fails to supply...