Data Agent

AdithyaSK 's Collections

updated about 2 hours ago

Sleeping

1

Harbor Eval Visualizations

📊

1

Explore model pass@k results with interactive dashboard
AdithyaSK/data_agent_rl_environment_train

Updated Jun 18 • 295

Note Canonical TRAINING suite: 2238 Harbor data-analysis tasks over real Kaggle datasets. Single correctness reward.
AdithyaSK/data_agent_rl_environment_eval

Updated Jun 11 • 110 • 1

Note Held-out EVAL suite: 366 verified Harbor tasks. For benchmarking, not training.
AdithyaSK/data_agent_rl

Viewer • Updated May 19 • 29.5k • 26

Note Source-of-truth row-level split manifest (parquet, eval/train, ~30k rows). The Harbor task suites are built from this.
AdithyaSK/jupyter-agent-kaggle-all

127 GB

Note Shared Kaggle DATA BUCKET every task pulls from at container start (keyed by BUCKET_PREFIX).
AdithyaSK/data_agent_rl_environment_train_subset_100

Updated Jun 18 • 156 • 2

Note Fast-iteration SUBSET: 100 easy (L1) numeric-reward tasks sampled from the train suite.
AdithyaSK/data_agent_rl_environment_train_multireward

Updated 22 days ago • 2.62k • 1

Note Train suite (2238) with a 3-REWARD verifier: correctness + submission + tool_efficiency (reward.json).
AdithyaSK/data_agent_rl_environment_train_difficulty_ranked

Updated 5 days ago • 45

Note Train suite (2238) ORDERED easy->hard by empirical pass@4 difficulty; registry carries rank/difficulty/solve_frac. Curriculum-ready + multireward.