paper-with-me

홈 › Papers

Benchmarking Mythos-Linked Bug Rediscovery

2026-05-17 · Isaac David, Arthur Gervais arxiv

Anthropic's April 2026 Mythos materials combine benchmark claims with concrete bug-finding stories across OpenBSD, FreeBSD, Linux, FFmpeg, and browsers. This paper reports a controlled target-file rediscovery experiment on six public or high-confidence Mythos-linked systems tasks. Each model receives the same target file or files, read-only source tools, three repeats per task, and one manual target-matching rubric; prompts omit CVE identifiers, patch hashes, advisory text, author names, disclosure dates, and answer key root cause language. The experiment contains 54 counted model-task attempts: three models, six tasks, and three repeats, giving 18 attempts per model. GPT-5.5 xhigh achieves 5/18 target rediscoveries, covering 2/6 tasks; counting one wrong-target mpegts.c finding separately gives 3/6 distinct core bugs. Claude Opus 4.7 achieves 1/18 target rediscoveries, covering 1/6 tasks. Kimi K2 records 0/18 target rediscoveries. The dominant failure mode is early commitment to plausible alternate candidates within the assigned file: models often submit source-grounded hypotheses while missing the specific invariant corrected by public Mythos patch evidence. These results do not refute Anthropic's undisclosed workflow, but show that under this favorable target-file scaffold, systems-specific prompting yields only six target matches across 54 counted attempts.

📄 PDF Abstract BibTeX arXiv:2605.17416

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Fast Symbolic Regression Benchmarking

2025-08-20 · Viktor Martinek arxiv

Symbolic regression (SR) uncovers mathematical models from data. Several benchmarks have been proposed to compare the performance of SR algorithms. However, existing ground-truth rediscovery benchmarks overemphasize the …

Mythos and the Unverified Cage: Z3-Based Pre-Deployment Verification for Frontier-Model Sandbox Infrastructure

2026-04-22 · Dominik Blain arxiv

The April 2026 Claude Mythos sandbox escape exposed a critical weakness in frontier AI containment: the infrastructure surrounding advanced models remains susceptible to formally characterizable arithmetic vulnerabilitie…

The Rediscovery Hypothesis: Language Models Need to Meet Linguistics

2021-03-02 · Vassilina Nikoulina, Maxat Tezekbayev, Nuradil Kozhakhmet, Madina Babazhanova 외

There is an ongoing debate in the NLP community whether modern language models contain linguistic knowledge, recovered through so-called probes. In this paper, we study whether linguistic knowledge is a necessary conditi…

Language ModelingLanguage Modelling

FIRE-Bench: Evaluating Agents on the Rediscovery of Scientific Insights

2026-02-02 · Zhen Wang, Fan Bai, Zhongyan Luo, Jinyan Su 외 arxiv

Autonomous agents powered by large language models (LLMs) promise to accelerate scientific discovery end-to-end, but rigorously evaluating their capacity for verifiable discovery remains a central challenge. Existing ben…

Evaluating whether AI models would sabotage AI safety research

2026-04-27 · Robert Kirk, Alexandra Souly, Kai Fronsdal, Abby D'Cruz 외 arxiv

We evaluate the propensity of frontier models to sabotage or refuse to assist with safety research when deployed as AI research agents within a frontier AI company. We apply two complementary evaluations to four Claude m…