Datasets:

yannabadie
/

sage-topology-dataset

text stringlengths 1 120
step:0 mean= max= entropy= resp= VAL=0.2213870248141197
step:1
1
64
64 mean=0.2170059084892273 max=1.0217115879058838 entropy=0.05569755600299686 resp=335.984375
step:2
2
63
63 mean=0.22343748807907104 max=0.699999988079071 entropy=0.05957562697585672 resp=356.84375
step:3
3
65
65 mean=0.21054688096046448 max=0.699999988079071 entropy=0.06499485007952899 resp=422.40625
step:4
4
63
63 mean=0.22829081118106842 max=1.0212221145629883 entropy=0.06146638287464157 resp=407.53125
step:5
5
63
63 mean=0.19544270634651184 max=0.22499999403953552 entropy=0.0619709171296563 resp=342.359375
step:6
6
66
66 mean=0.2085937261581421 max=0.699999988079071 entropy=0.059504195814952254 resp=415.171875
step:7
7
63
63 mean=0.2376192808151245 max=1.0270546674728394 entropy=0.061034475336782634 resp=363.5625
step:8
8
62
62 mean=0.20911458134651184 max=0.699999988079071 entropy=0.06103689764859155 resp=423.421875
step:9
9
62
62 mean=0.22604167461395264 max=0.699999988079071 entropy=0.05910565488738939 resp=349.296875
step:10
10
64
64 mean=0.20781250298023224 max=0.699999988079071 entropy=0.06700595148140565 resp=455.375
step:11
11
64
64 mean=0.21838031709194183 max=1.0180070400238037 entropy=0.06005544337676838 resp=420.859375
step:12
12
62
62 mean=0.21465450525283813 max=1.0212221145629883 entropy=0.05643557256553322 resp=376.40625
step:13
13
63
63 mean=0.23836831748485565 max=1.0599726438522339 entropy=0.0661239794571884 resp=390.0625
step:14
14
65
65 mean=0.22356770932674408 max=0.699999988079071 entropy=0.06507335626520216 resp=511.328125
step:15
15
62
62 mean=0.2174498587846756 max=1.0215673446655273 entropy=0.06258630874799564 resp=400.6875
step:16
16
63
63 mean=0.21135373413562775 max=1.0266392230987549 entropy=0.07695581967709586 resp=477.328125
step:17
17
64
64 mean=0.29979208111763 max=1.0266392230987549 entropy=0.06534311379073188 resp=424.328125
step:18
18
63
63 mean=0.21067708730697632 max=0.699999988079071 entropy=0.07384027025545947 resp=479.203125
step:19
19
62
62 mean=0.23586231470108032 max=1.0212221145629883 entropy=0.07421565405093133 resp=422.09375
step:20
20
62
62 mean=0.19908854365348816 max=0.699999988079071 entropy=0.05134229137911461 resp=375.140625
step:21
21
63
63 mean=0.25410765409469604 max=1.0212221145629883 entropy=0.06597896944731474 resp=413.40625
step:22
22
62
62 mean=0.20533853769302368 max=0.22499999403953552 entropy=0.07007976801833138 resp=412.15625
step:23
23
63
63 mean=0.2131413221359253 max=1.0243785381317139 entropy=0.06993384304223582 resp=496.640625
step:24
24
62
62 mean=0.2211703360080719 max=1.0215673446655273 entropy=0.057703598926309496 resp=357.6875
step:25
25
63

End of preview. Expand in Data Studio

SAGE Topology Training Dataset

Training data for Nemotron-Orchestrator-8B topology generation.

Formats

File	Format	Entries	Description
`verl_topology_train_json.parquet`	JSON	12,303	Full training set (JSON tool-calls)
`verl_topology_curated_json.parquet`	JSON	498	Curated subset (adaptive topologies)
`verl_topology_train.parquet`	YAML (legacy)	12,303	Original YAML format

Sources

GSM8K, BigCodeBench, APPS, CodeContests
GPT-5.4 generated topologies (audited)
Human-curated adaptive examples

Downloads last month: 21