grg commited on
Commit
2be5d2e
1 Parent(s): 9ae2a55

Adding hermes_3_llama_3.1_8b

Browse files
static/leaderboard.csv CHANGED
@@ -1,34 +1,35 @@
1
  Model,Ordinal (Win rate),Cardinal (Score),RO Stability,Stress,CFI,SRMR,RMSEA
2
- gemma-2-2b-it,0.3732638888888889,0.3309858600428668,0.14746606707946294,0.263080165752695,0.40932500000000005,0.550475,0.5377472222222222
3
- gemma-2-9b-it,0.7387152777777778,0.6020857503693501,0.43782539244147833,0.20116278903333318,0.7543666666666667,0.23989444444444452,0.24792499999999995
4
- gemma-2-27b-it,0.6319444444444444,0.5270946699366518,0.3917304045417486,0.2058170364515589,0.5997861111111111,0.37121111111111116,0.37292222222222227
5
- phi-3-mini-128k-instruct,0.3246527777777778,0.32984992817164005,0.039299993295009855,0.281800547806919,0.5861361111111111,0.42524166666666674,0.3974944444444444
6
- phi-3-medium-128k-instruct,0.3307291666666667,0.30802986933853177,0.09692037989916814,0.2651981204439735,0.43025555555555556,0.5503277777777777,0.5381722222222222
7
- phi-3.5-mini-instruct,0.24348958333333334,0.2680653144619754,0.0361229186530762,0.28422749224983457,0.40715555555555555,0.5721138888888888,0.5507833333333333
8
- phi-3.5-MoE-instruct,0.3962673611111111,0.36128192067041315,0.10985291697837646,0.2739229692168671,0.5530944444444444,0.4248777777777778,0.40345
9
- Mistral-7B-Instruct-v0.1,0.21961805555555555,0.26609566354811315,0.027216280472015988,0.2829498135031582,0.38917777777777773,0.5561138888888888,0.530213888888889
10
- Mistral-7B-Instruct-v0.2,0.3506944444444444,0.32133832899241477,0.14417876497818388,0.265188983528973,0.3802722222222222,0.5727305555555555,0.5483611111111111
11
- Mistral-7B-Instruct-v0.3,0.2569444444444444,0.26572479479146804,0.07960539866974455,0.2742399030139009,0.31385,0.6241,0.6081333333333333
12
- Mixtral-8x7B-Instruct-v0.1,0.4431423611111111,0.3819009850972602,0.21473356319081474,0.2624402608740656,0.45275,0.5034666666666667,0.4905694444444444
13
- Mixtral-8x22B-Instruct-v0.1,0.3381076388888889,0.31529864972153404,0.1414001940345544,0.2548838005881672,0.3772361111111111,0.5810888888888889,0.5844750000000001
14
- command_r_plus,0.5828993055555556,0.4995356672762356,0.3429686514651868,0.23811982320641845,0.6033000000000001,0.3740166666666668,0.3667527777777777
15
- llama_3_8b_instruct,0.4943576388888889,0.4295836112681494,0.24527785038654715,0.245806400289881,0.5498222222222222,0.42656388888888896,0.42189444444444446
16
- llama_3_70b_instruct,0.7855902777777778,0.6839540364836003,0.607020698814379,0.18525883672204868,0.7210055555555557,0.2346083333333333,0.25758888888888887
17
- llama_3.1_8b_instruct,0.5737847222222222,0.4786874422110324,0.4295080949846363,0.22060228669473025,0.4305722222222223,0.5455027777777777,0.553
18
- llama_3.1_70b_instruct,0.8272569444444444,0.7172545013390067,0.691365862744007,0.1709718847084183,0.6979472222222223,0.2636777777777777,0.2907250000000001
19
- llama_3.1_405b_instruct_4bit,0.7404513888888888,0.6490864350383405,0.7232098126552619,0.1702199925365422,0.4875722222222223,0.4963444444444445,0.5211555555555556
20
- llama_3.2_1b_instruct,0.2152777777777778,0.2522036562381785,0.027192115495770382,0.29255310096654275,0.37450000000000006,0.5990222222222223,0.5740638888888888
21
- llama_3.2_3b_instruct,0.4001736111111111,0.3615804465210719,0.13450325180647235,0.27485276839064654,0.5017,0.44956666666666667,0.4226500000000001
22
- Qwen2-7B-Instruct,0.4123263888888889,0.36370005127542027,0.25108519506513916,0.25776537005719313,0.3560861111111111,0.6009722222222222,0.5920888888888889
23
- Qwen2-72B-Instruct,0.5802951388888888,0.5461212335522644,0.6465993243020925,0.20297742879025626,0.3045,0.6543138888888889,0.6646361111111111
24
- Qwen2.5-0.5B-Instruct,0.2899305555555556,0.3005554090516966,0.002970456550606876,0.2928913315666324,0.5371250000000001,0.44709722222222226,0.404575
25
- Qwen2.5-7B-Instruct,0.6163194444444444,0.5163098181421168,0.333554494486959,0.2505866550331236,0.6473694444444444,0.30400277777777773,0.29651944444444434
26
- Qwen2.5-32B-Instruct,0.7408854166666666,0.656917654644944,0.6724190751477237,0.1806656189868978,0.5603222222222223,0.40237500000000004,0.41161666666666663
27
- Qwen2.5-72B-Instruct,0.8276909722222222,0.7104489147495714,0.6974116787371809,0.16176650806326276,0.6734583333333333,0.2993,0.3184472222222223
28
- gpt-3.5-turbo-0125,0.24609375,0.28218378886707396,0.08240359836763214,0.28728574920060357,0.3873055555555555,0.599925,0.572238888888889
29
- gpt-4o-0513,0.6948784722222222,0.5989532974661671,0.5122163952167618,0.19201420113771173,0.6235416666666667,0.34458611111111115,0.3441805555555555
30
- gpt-4o-mini-2024-07-18,0.37890625,0.3418785071827972,0.13575309046266867,0.2707065266105181,0.44214722222222214,0.5004583333333332,0.47896666666666665
31
- Mistral-Large-Instruct-2407,0.8467881944444444,0.7374229691535793,0.7644582301049158,0.16944638941325085,0.6510750000000001,0.31028611111111104,0.3297916666666667
32
- Mistral-Nemo-Instruct-2407,0.5876736111111112,0.5262426956484347,0.4414072595011627,0.21142636170606344,0.5161,0.42923055555555545,0.43113055555555546
33
- Mistral-Small-Instruct-2409,0.7782118055555556,0.6890378862258165,0.6416815833333804,0.1894343546381,0.6840472222222221,0.2601583333333335,0.2888777777777778
34
- dummy,0.18880208333333334,0.2291015386716794,-0.009004148398032956,0.2928877637010999,0.3755222222222222,0.622275,0.5915305555555557
 
 
1
  Model,Ordinal (Win rate),Cardinal (Score),RO Stability,Stress,CFI,SRMR,RMSEA
2
+ hermes_3_llama_3.1_8b_instruct,0.4638047138047138,0.4117221943281449,0.16520527634373441,0.25311021741644446,0.5822944444444444,0.3526722222222223,0.3435555555555555
3
+ gemma-2-2b-it,0.3733164983164983,0.3309858600428668,0.14746606707946294,0.263080165752695,0.40932500000000005,0.550475,0.5377472222222222
4
+ gemma-2-9b-it,0.7432659932659933,0.6020857503693501,0.43782539244147833,0.20116278903333318,0.7543666666666667,0.23989444444444452,0.24792499999999995
5
+ gemma-2-27b-it,0.6372053872053871,0.5270946699366518,0.3917304045417486,0.2058170364515589,0.5997861111111111,0.37121111111111116,0.37292222222222227
6
+ phi-3-mini-128k-instruct,0.3253367003367003,0.32984992817164005,0.039299993295009855,0.281800547806919,0.5861361111111111,0.42524166666666674,0.3974944444444444
7
+ phi-3-medium-128k-instruct,0.3282828282828283,0.30802986933853177,0.09692037989916814,0.2651981204439735,0.43025555555555556,0.5503277777777777,0.5381722222222222
8
+ phi-3.5-mini-instruct,0.24200336700336697,0.2680653144619754,0.0361229186530762,0.28422749224983457,0.40715555555555555,0.5721138888888888,0.5507833333333333
9
+ phi-3.5-MoE-instruct,0.39520202020202017,0.36128192067041315,0.10985291697837646,0.2739229692168671,0.5530944444444444,0.4248777777777778,0.40345
10
+ Mistral-7B-Instruct-v0.1,0.2175925925925926,0.26609566354811315,0.027216280472015988,0.2829498135031582,0.38917777777777773,0.5561138888888888,0.530213888888889
11
+ Mistral-7B-Instruct-v0.2,0.35058922558922556,0.32133832899241477,0.14417876497818388,0.265188983528973,0.3802722222222222,0.5727305555555555,0.5483611111111111
12
+ Mistral-7B-Instruct-v0.3,0.2558922558922559,0.26572479479146804,0.07960539866974455,0.2742399030139009,0.31385,0.6241,0.6081333333333333
13
+ Mixtral-8x7B-Instruct-v0.1,0.44486531986531985,0.3819009850972602,0.21473356319081474,0.2624402608740656,0.45275,0.5034666666666667,0.4905694444444444
14
+ Mixtral-8x22B-Instruct-v0.1,0.3421717171717171,0.31529864972153404,0.1414001940345544,0.2548838005881672,0.3772361111111111,0.5810888888888889,0.5844750000000001
15
+ command_r_plus,0.5892255892255892,0.4995356672762356,0.3429686514651868,0.23811982320641845,0.6033000000000001,0.3740166666666668,0.3667527777777777
16
+ llama_3_8b_instruct,0.49915824915824913,0.4295836112681494,0.24527785038654715,0.245806400289881,0.5498222222222222,0.42656388888888896,0.42189444444444446
17
+ llama_3_70b_instruct,0.7882996632996634,0.6839540364836003,0.607020698814379,0.18525883672204868,0.7210055555555557,0.2346083333333333,0.25758888888888887
18
+ llama_3.1_8b_instruct,0.5795454545454546,0.4786874422110324,0.4295080949846363,0.22060228669473025,0.4305722222222223,0.5455027777777777,0.553
19
+ llama_3.1_70b_instruct,0.8282828282828283,0.7172545013390067,0.691365862744007,0.1709718847084183,0.6979472222222223,0.2636777777777777,0.2907250000000001
20
+ llama_3.1_405b_instruct_4bit,0.7394781144781145,0.6490864350383405,0.7232098126552619,0.1702199925365422,0.4875722222222223,0.4963444444444445,0.5211555555555556
21
+ llama_3.2_1b_instruct,0.21506734006734007,0.2522036562381785,0.027192115495770382,0.29255310096654275,0.37450000000000006,0.5990222222222223,0.5740638888888888
22
+ llama_3.2_3b_instruct,0.3985690235690236,0.3615804465210719,0.13450325180647235,0.27485276839064654,0.5017,0.44956666666666667,0.4226500000000001
23
+ Qwen2-7B-Instruct,0.41498316498316495,0.36370005127542027,0.25108519506513916,0.25776537005719313,0.3560861111111111,0.6009722222222222,0.5920888888888889
24
+ Qwen2-72B-Instruct,0.5829124579124579,0.5461212335522644,0.6465993243020925,0.20297742879025626,0.3045,0.6543138888888889,0.6646361111111111
25
+ Qwen2.5-0.5B-Instruct,0.2882996632996633,0.3005554090516966,0.002970456550606876,0.2928913315666324,0.5371250000000001,0.44709722222222226,0.404575
26
+ Qwen2.5-7B-Instruct,0.622053872053872,0.5163098181421168,0.333554494486959,0.2505866550331236,0.6473694444444444,0.30400277777777773,0.29651944444444434
27
+ Qwen2.5-32B-Instruct,0.7411616161616161,0.656917654644944,0.6724190751477237,0.1806656189868978,0.5603222222222223,0.40237500000000004,0.41161666666666663
28
+ Qwen2.5-72B-Instruct,0.8274410774410774,0.7104489147495714,0.6974116787371809,0.16176650806326276,0.6734583333333333,0.2993,0.3184472222222223
29
+ gpt-3.5-turbo-0125,0.24452861952861948,0.28218378886707396,0.08240359836763214,0.28728574920060357,0.3873055555555555,0.599925,0.572238888888889
30
+ gpt-4o-0513,0.6973905723905723,0.5989532974661671,0.5122163952167618,0.19201420113771173,0.6235416666666667,0.34458611111111115,0.3441805555555555
31
+ gpt-4o-mini-2024-07-18,0.37542087542087543,0.3418785071827972,0.13575309046266867,0.2707065266105181,0.44214722222222214,0.5004583333333332,0.47896666666666665
32
+ Mistral-Large-Instruct-2407,0.8455387205387205,0.7374229691535793,0.7644582301049158,0.16944638941325085,0.6510750000000001,0.31028611111111104,0.3297916666666667
33
+ Mistral-Nemo-Instruct-2407,0.5904882154882155,0.5262426956484347,0.4414072595011627,0.21142636170606344,0.5161,0.42923055555555545,0.43113055555555546
34
+ Mistral-Small-Instruct-2409,0.7815656565656567,0.6890378862258165,0.6416815833333804,0.1894343546381,0.6840472222222221,0.2601583333333335,0.2888777777777778
35
+ dummy,0.18855218855218855,0.2291015386716794,-0.009004148398032956,0.2928877637010999,0.3755222222222222,0.622275,0.5915305555555557
static/models_data/cardinal.svg CHANGED
static/models_data/gemma-2-9b-it/cfa_metrics.csv CHANGED
@@ -1,10 +1,3 @@
1
  Context chunk,CFI,TLI,SRMR,RMSEA
2
- chunk_0,0.888425,0.8612500000000001,0.094,0.0795
3
- chunk_1,0.897575,0.879875,0.09025,0.0885
4
- chunk_2,0.699975,0.6807000000000001,0.3105,0.3014
5
- chunk_3,0.673925,0.657025,0.31282499999999996,0.307475
6
- chunk_4,0.47187500000000004,0.45735000000000003,0.5402499999999999,0.5511
7
  chunk_chess_0,0.911425,0.8817499999999999,0.08405,0.1146
8
  chunk_grammar_1,0.901925,0.87,0.082675,0.110675
9
- chunk_no_conv,0.702225,0.683525,0.31505000000000005,0.33699999999999997
10
- chunk_svs_no_conv,0.64195,0.610225,0.32944999999999997,0.341075
 
1
  Context chunk,CFI,TLI,SRMR,RMSEA
 
 
 
 
 
2
  chunk_chess_0,0.911425,0.8817499999999999,0.08405,0.1146
3
  chunk_grammar_1,0.901925,0.87,0.082675,0.110675
 
 
static/models_data/hermes_3_llama_3.1_8b/cfa_metrics.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ Context chunk,CFI,TLI,SRMR,RMSEA
2
+ chunk_0,0.5613,0.49920000000000003,0.337775,0.326675
3
+ chunk_1,0.64515,0.6025,0.323025,0.30965000000000004
4
+ chunk_2,0.359575,0.318875,0.555975,0.546925
5
+ chunk_3,0.825825,0.85975,0.0964,0.074575
6
+ chunk_4,0.655175,0.6530750000000001,0.32227500000000003,0.297125
7
+ chunk_chess_0,0.8679250000000001,0.8399749999999999,0.097225,0.091075
8
+ chunk_grammar_1,0.44962500000000005,0.349275,0.341025,0.33925
9
+ chunk_no_conv,0.6267,0.590625,0.3256,0.34885
10
+ chunk_svs_no_conv,0.249375,0.2491,0.77475,0.757875
static/models_data/hermes_3_llama_3.1_8b/matrix.svg ADDED
static/models_data/hermes_3_llama_3.1_8b/model_detail.html ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ <p>
2
+ This open-source model was created by <a target="_blank" href="https://nousresearch.com/">Nous Research</a>.
3
+ You can find the release blog post <a target="_blank" href="https://nousresearch.com/hermes3/">here</a>.
4
+ The model is available on the huggingface hub: <a target="_blank" href="https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B">https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B</a>.
5
+ The 8B model was supports up to 128K token contexts.
6
+ </p>
static/models_data/hermes_3_llama_3.1_8b/ranks.svg ADDED
static/models_data/hermes_3_llama_3.1_8b/structure.svg ADDED
static/models_data/ordinal.svg CHANGED