Json604 / sycophancy-bench Star 0 Code Issues Pull requests Do models catch flawed ML results? Flawed reports with byte-identical matched controls. benchmark machine-learning research-integrity evaluation-methodology llm-evaluation sycophancy llm-judge agent-evaluation evaluation-harness contamination-free Updated Aug 20, 2026 Python
NagaYu / ingot Star 0 Code Issues Pull requests Inverted generation: construct the label first, render the sentence second. Strictly-labelled, contamination-free synthetic data without a judge. nlp icalendar japanese dataset-generation rrule synthetic-data constrained-decoding small-language-models contamination-free inverted-generation Updated Aug 23, 2026 Python
Json604 / rewardhack-bench Star 0 Code Issues Pull requests Predict how a policy games a reward. RL environments scored by an executable-exploit verifier. benchmark machine-learning reinforcement-learning ai-alignment llm-evaluation agent-evaluation reward-hacking evaluation-harness specification-gaming contamination-free Updated Aug 20, 2026 Python
Json604 / silentbug-bench Star 0 Code Issues Pull requests ML training defects that never crash. An agentic benchmark with patch-and-rerun, unit-test verification. debugging benchmark machine-learning pytorch ml-engineering llm-evaluation agentic-ai agent-evaluation evaluation-harness contamination-free Updated Aug 20, 2026 Python
LiveMacroEval / LiveMacroEval Star 0 Code Issues Pull requests A live, contamination-resistant benchmark for evaluating LLM agents on U.S. macroeconomic nowcasting. agent finance economics nowcasting llm-evaluation llm-agents contamination-free Updated Aug 29, 2026 Python