saketh-chervu's picture
Upload eval_data/meta.json with huggingface_hub
db4b388 verified
Raw
History Blame Contribute Delete
5.05 kB
{
"subexp": "s3_str2math",
"variant": "cot",
"arm": "correct",
"seed": 0,
"atoms": [
"swap_pairs",
"reverse_first_half",
"interleave_halves",
"nice_chars",
"charset",
"multiply",
"line_intercepts",
"rect_area_perimeter",
"gcf",
"lcm",
"venn"
],
"train_comps": [
"swap_pairs_o_interleave_halves",
"interleave_halves_o_swap_pairs"
],
"heldout_comps": [
"multiply_o_gcf",
"gcf_o_multiply",
"multiply_o_lcm",
"lcm_o_multiply",
"gcf_o_lcm",
"lcm_o_gcf",
"multiply_o_venn",
"gcf_o_venn",
"lcm_o_venn"
],
"compositions": {
"swap_pairs_o_interleave_halves": {
"role": "train",
"cross": "ss",
"n_train": 200,
"n_seen_eval": 200
},
"interleave_halves_o_swap_pairs": {
"role": "train",
"cross": "ss",
"n_train": 200,
"n_seen_eval": 200
},
"multiply_o_gcf": {
"role": "heldout",
"cross": "mm",
"n_eval": 200,
"distinct_answers": 195,
"top_answer_share": 0.01
},
"gcf_o_multiply": {
"role": "heldout",
"cross": "mm",
"n_eval": 200,
"distinct_answers": 173,
"top_answer_share": 0.02
},
"multiply_o_lcm": {
"role": "heldout",
"cross": "mm",
"n_eval": 200,
"distinct_answers": 176,
"top_answer_share": 0.015
},
"lcm_o_multiply": {
"role": "heldout",
"cross": "mm",
"n_eval": 200,
"distinct_answers": 175,
"top_answer_share": 0.02
},
"gcf_o_lcm": {
"role": "heldout",
"cross": "mm",
"n_eval": 200,
"distinct_answers": 169,
"top_answer_share": 0.025
},
"lcm_o_gcf": {
"role": "heldout",
"cross": "mm",
"n_eval": 200,
"distinct_answers": 195,
"top_answer_share": 0.015
},
"multiply_o_venn": {
"role": "heldout",
"cross": "mm",
"n_eval": 200,
"distinct_answers": 192,
"top_answer_share": 0.01
},
"gcf_o_venn": {
"role": "heldout",
"cross": "mm",
"n_eval": 200,
"distinct_answers": 186,
"top_answer_share": 0.015
},
"lcm_o_venn": {
"role": "heldout",
"cross": "mm",
"n_eval": 200,
"distinct_answers": 191,
"top_answer_share": 0.015
}
},
"train_heldout_chain_overlap": [],
"heldout_instance_collisions": 0,
"digit_mix": {
"swap_pairs": {
"n": 800,
"digit4": 199,
"frac": 0.249
},
"reverse_first_half": {
"n": 800,
"digit4": 203,
"frac": 0.254
},
"interleave_halves": {
"n": 800,
"digit4": 196,
"frac": 0.245
}
},
"collisions_total": 0,
"degeneracy": {
"swap_pairs_o_interleave_halves": {
"cross": "ss",
"pass": true,
"detail": {
"letters": {
"order_sensitive": true,
"differs_from_f": true,
"differs_from_g": true
},
"digit4": {
"order_sensitive": true,
"differs_from_f": true,
"differs_from_g": true
}
}
},
"interleave_halves_o_swap_pairs": {
"cross": "ss",
"pass": true,
"detail": {
"letters": {
"order_sensitive": true,
"differs_from_f": true,
"differs_from_g": true
},
"digit4": {
"order_sensitive": true,
"differs_from_f": true,
"differs_from_g": true
}
}
},
"multiply_o_gcf": {
"cross": "mm",
"pass": true,
"outer_changes_handoff_frac": 1.0,
"n_generated": 40,
"note": "non-commuting ops"
},
"gcf_o_multiply": {
"cross": "mm",
"pass": true,
"outer_changes_handoff_frac": 1.0,
"n_generated": 40,
"note": "non-commuting ops"
},
"multiply_o_lcm": {
"cross": "mm",
"pass": true,
"outer_changes_handoff_frac": 1.0,
"n_generated": 40,
"note": "non-commuting ops"
},
"lcm_o_multiply": {
"cross": "mm",
"pass": true,
"outer_changes_handoff_frac": 1.0,
"n_generated": 40,
"note": "non-commuting ops"
},
"gcf_o_lcm": {
"cross": "mm",
"pass": true,
"outer_changes_handoff_frac": 1.0,
"n_generated": 40,
"note": "non-commuting ops"
},
"lcm_o_gcf": {
"cross": "mm",
"pass": true,
"outer_changes_handoff_frac": 1.0,
"n_generated": 40,
"note": "non-commuting ops"
},
"multiply_o_venn": {
"cross": "mm",
"pass": true,
"outer_changes_handoff_frac": 1.0,
"n_generated": 40,
"note": "venn inner-only"
},
"gcf_o_venn": {
"cross": "mm",
"pass": true,
"outer_changes_handoff_frac": 1.0,
"n_generated": 40,
"note": "venn inner-only"
},
"lcm_o_venn": {
"cross": "mm",
"pass": true,
"outer_changes_handoff_frac": 1.0,
"n_generated": 40,
"note": "venn inner-only"
}
},
"degeneracy_fails": []
}