formal-mas / data /retrieval_eval /results_dev.json
NicolasT08's picture
Formal-MAS source code
dfe1826 verified
Raw History Blame Contribute Delete
5.36 kB
{
"created_utc": "2026-10-19T23:43:08+00:00",
"split": "dev",
"n": 134,
"ks": [
4,
8,
16,
32
],
"depth": 100,
"corpus": "corpus_eval.jsonl",
"n_generalized": 40,
"results": {
"bm25": {
"c": {
"R@4": 0.43283582089552236,
"R@8": 0.5074626865671642,
"R@16": 0.6343283582089553,
"R@32": 0.7238805970149254,
"MRR": 0.3599456654442559
},
"b": {
"R@4": 0.44776119402985076,
"R@8": 0.5373134328358209,
"R@16": 0.6716417910447762,
"R@32": 0.7611940298507462,
"MRR": 0.3683568684413515
}
},
"bge-base/no_instr": {
"c": {
"R@4": 0.26865671641791045,
"R@8": 0.3283582089552239,
"R@16": 0.41044776119402987,
"R@32": 0.5223880597014925,
"MRR": 0.20708555432556183
},
"b": {
"R@4": 0.2835820895522388,
"R@8": 0.3582089552238806,
"R@16": 0.4626865671641791,
"R@32": 0.5746268656716418,
"MRR": 0.21746678884009404
}
},
"bge-base/instr": {
"c": {
"R@4": 0.30597014925373134,
"R@8": 0.39552238805970147,
"R@16": 0.5149253731343284,
"R@32": 0.6194029850746269,
"MRR": 0.24694153308866407
},
"b": {
"R@4": 0.31343283582089554,
"R@8": 0.44029850746268656,
"R@16": 0.5597014925373134,
"R@32": 0.6716417910447762,
"MRR": 0.2564905098666834
}
},
"rrf(bm25+bge-base/no_instr)": {
"c": {
"R@4": 0.3283582089552239,
"R@8": 0.41044776119402987,
"R@16": 0.5970149253731343,
"R@32": 0.7089552238805971,
"MRR": 0.27157358376374574
},
"b": {
"R@4": 0.3358208955223881,
"R@8": 0.4552238805970149,
"R@16": 0.6343283582089553,
"R@32": 0.753731343283582,
"MRR": 0.28169675330855704
}
},
"rrf(bm25+bge-base/instr)": {
"c": {
"R@4": 0.373134328358209,
"R@8": 0.5,
"R@16": 0.6194029850746269,
"R@32": 0.7611940298507462,
"MRR": 0.323935069392284
},
"b": {
"R@4": 0.3880597014925373,
"R@8": 0.5447761194029851,
"R@16": 0.6716417910447762,
"R@32": 0.7985074626865671,
"MRR": 0.3350275796941803
}
},
"bge-m3/no_instr": {
"c": {
"R@4": 0.3656716417910448,
"R@8": 0.4552238805970149,
"R@16": 0.5522388059701493,
"R@32": 0.6567164179104478,
"MRR": 0.2951335548669053
},
"b": {
"R@4": 0.39552238805970147,
"R@8": 0.5,
"R@16": 0.5970149253731343,
"R@32": 0.6940298507462687,
"MRR": 0.3149321866706431
}
},
"rrf(bm25+bge-m3/no_instr)": {
"c": {
"R@4": 0.41044776119402987,
"R@8": 0.4925373134328358,
"R@16": 0.5970149253731343,
"R@32": 0.753731343283582,
"MRR": 0.3399236750140224
},
"b": {
"R@4": 0.43283582089552236,
"R@8": 0.5298507462686567,
"R@16": 0.6417910447761194,
"R@32": 0.7835820895522388,
"MRR": 0.3539497532920119
}
},
"e5-base/no_instr": {
"c": {
"R@4": 0.3582089552238806,
"R@8": 0.4626865671641791,
"R@16": 0.5447761194029851,
"R@32": 0.664179104477612,
"MRR": 0.27693185026669315
},
"b": {
"R@4": 0.3880597014925373,
"R@8": 0.5223880597014925,
"R@16": 0.6044776119402985,
"R@32": 0.7164179104477612,
"MRR": 0.2905748261517756
}
},
"e5-base/instr": {
"c": {
"R@4": 0.3656716417910448,
"R@8": 0.44029850746268656,
"R@16": 0.5447761194029851,
"R@32": 0.6492537313432836,
"MRR": 0.2699728290275252
},
"b": {
"R@4": 0.39552238805970147,
"R@8": 0.47761194029850745,
"R@16": 0.5970149253731343,
"R@32": 0.6940298507462687,
"MRR": 0.29035977733918816
}
},
"rrf(bm25+e5-base/no_instr)": {
"c": {
"R@4": 0.417910447761194,
"R@8": 0.4925373134328358,
"R@16": 0.5895522388059702,
"R@32": 0.7388059701492538,
"MRR": 0.3280031737655239
},
"b": {
"R@4": 0.43283582089552236,
"R@8": 0.5298507462686567,
"R@16": 0.6343283582089553,
"R@32": 0.7835820895522388,
"MRR": 0.33893815046023923
}
},
"rrf(bm25+e5-base/instr)": {
"c": {
"R@4": 0.43283582089552236,
"R@8": 0.5,
"R@16": 0.6343283582089553,
"R@32": 0.7611940298507462,
"MRR": 0.31906612553566926
},
"b": {
"R@4": 0.44776119402985076,
"R@8": 0.5373134328358209,
"R@16": 0.6716417910447762,
"R@32": 0.7985074626865671,
"MRR": 0.327167222053173
}
},
"rrf(bm25+e5-base/instr)+generic": {
"c": {
"R@4": 0.43283582089552236,
"R@8": 0.5,
"R@16": 0.6417910447761194,
"R@32": 0.7611940298507462,
"MRR": 0.32262696168114985
},
"b": {
"R@4": 0.44776119402985076,
"R@8": 0.5373134328358209,
"R@16": 0.6791044776119403,
"R@32": 0.7985074626865671,
"MRR": 0.3307902472533799
}
}
}
}