Decomposing and measuring evaluation awareness in existing benchmarks and our proposed EvalAwareBench.
-
Updated
Jun 1, 2026 - Python
#
evaluation-awareness
Here are 7 public repositories matching this topic...
Repository for the arXiv 2026 prepring "Models That Know How Evaluations Are Designed Score Safer"
-
Updated
Aug 17, 2026 - Jupyter Notebook
Synthetic-document fine-tuning on Qwen2.5-7B: a controlled study of whether SDF installs sandbagging, finding a layered recognition/generation/behavior dissociation.
model-organisms ai-safety ai-alignment sandbagging evaluation-awareness synthetic-document-finetuning
-
Updated
May 25, 2026 - Python
Agentic evaluation of misalignment, scheming and evaluation awareness in LLM agents. Scenario text encoded at rest to resist training-data contamination.
python benchmark model-evaluation ai-safety ai-agents ai-alignment llm llm-evaluation scheming evaluation-awareness
-
Updated
Aug 31, 2026 - Python
Certified reachability of NLA-defined concepts: proving when a model can(not) recognize it is being tested. Mashes Anthropic's Natural Language Autoencoder with Raghunathan certified defenses, on GPT-2.
ai-safety interpretability ai-alignment gpt-2 adversarial-robustness certified-robustness mechanistic-interpretability evaluation-awareness
-
Updated
Jul 25, 2026 - Python
Testing which evaluation-context cues change Qwen3-32B's task-completion behavior.
-
Updated
Aug 31, 2026 - Python
[WIP] Code and datasets for the thesis "Sensitivity Analysis of Evaluation Awareness in Large Language Models" · Licenciatura en Ciencia de Datos, Universidad de Buenos Aires (UBA).
-
Updated
Aug 24, 2026 - Jupyter Notebook
Add this topic to your repo
To associate your repository with the evaluation-awareness topic, visit your repo's landing page and select "manage topics."