Hugging Face's logo

Datasets:
NPHardEval
/
NPHardEval-results

Modalities:
Text
Formats:
json
Size:
1K<n<10K
Tags:
License:
Libraries:
Datasets
Dask
mlcroissant
Dataset card Viewer Files Community
1
haoyang commited on
Commit
e09c49f
•
1 Parent(s): a24bc73

update params

Browse files
Files changed (13) hide show
  1. 01-ai/Yi-34B-Chat/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} +0 -0
  2. Claude-2/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} +1 -1
  3. Claude-Instant/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} +1 -1
  4. GPT-3.5-Turbo/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} +1 -1
  5. GPT-4-Turbo/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} +1 -1
  6. PaLM-2/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} +1 -1
  7. Qwen/Qwen-14B-Chat/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} +0 -0
  8. export.ipynb +15 -1
  9. lmsys/vicuna-13b-v1.3/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} +0 -0
  10. microsoft/phi-1_5/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} +0 -0
  11. microsoft/phi-2/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} +0 -0
  12. mistralai/Mistral-7B-Instruct-v0.1/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} +0 -0
  13. mosaicml/mpt-30b-instruct/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} +0 -0
01-ai/Yi-34B-Chat/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} RENAMED
File without changes
Claude-2/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} RENAMED
@@ -3,7 +3,7 @@
3
  "model_name": "Claude-2",
4
  "model_type": "close-source",
5
  "model_dtype": "?",
6
- "num_params": 0
7
  },
8
  "results": {
9
  "SAS": {
 
3
  "model_name": "Claude-2",
4
  "model_type": "close-source",
5
  "model_dtype": "?",
6
+ "num_params": 176
7
  },
8
  "results": {
9
  "SAS": {
Claude-Instant/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} RENAMED
@@ -3,7 +3,7 @@
3
  "model_name": "Claude-Instant",
4
  "model_type": "close-source",
5
  "model_dtype": "?",
6
- "num_params": 0
7
  },
8
  "results": {
9
  "SAS": {
 
3
  "model_name": "Claude-Instant",
4
  "model_type": "close-source",
5
  "model_dtype": "?",
6
+ "num_params": 60
7
  },
8
  "results": {
9
  "SAS": {
GPT-3.5-Turbo/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} RENAMED
@@ -3,7 +3,7 @@
3
  "model_name": "GPT-3.5-Turbo",
4
  "model_type": "close-source",
5
  "model_dtype": "?",
6
- "num_params": 0
7
  },
8
  "results": {
9
  "SAS": {
 
3
  "model_name": "GPT-3.5-Turbo",
4
  "model_type": "close-source",
5
  "model_dtype": "?",
6
+ "num_params": 175
7
  },
8
  "results": {
9
  "SAS": {
GPT-4-Turbo/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} RENAMED
@@ -3,7 +3,7 @@
3
  "model_name": "GPT-4-Turbo",
4
  "model_type": "close-source",
5
  "model_dtype": "?",
6
- "num_params": 0
7
  },
8
  "results": {
9
  "SAS": {
 
3
  "model_name": "GPT-4-Turbo",
4
  "model_type": "close-source",
5
  "model_dtype": "?",
6
+ "num_params": 1760
7
  },
8
  "results": {
9
  "SAS": {
PaLM-2/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} RENAMED
@@ -3,7 +3,7 @@
3
  "model_name": "PaLM-2",
4
  "model_type": "close-source",
5
  "model_dtype": "?",
6
- "num_params": 0
7
  },
8
  "results": {
9
  "SAS": {
 
3
  "model_name": "PaLM-2",
4
  "model_type": "close-source",
5
  "model_dtype": "?",
6
+ "num_params": 340
7
  },
8
  "results": {
9
  "SAS": {
Qwen/Qwen-14B-Chat/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} RENAMED
File without changes
export.ipynb CHANGED
@@ -46,6 +46,15 @@
46
  "new_df.model.unique()"
47
  ]
48
  },
 
 
 
 
 
 
 
 
 
49
  {
50
  "cell_type": "code",
51
  "execution_count": 4,
@@ -59,7 +68,7 @@
59
  " \"Phi-1.5\": \"microsoft/phi-1_5\",\n",
60
  " \"MPT-30b\": \"mosaicml/mpt-30b-instruct\",\n",
61
  " \"Phi-2\": \"microsoft/phi-2\",\n",
62
- " \"Qwen-14b\": \"Qwen/Qwen-14B-Chat\"\n",
63
  "}\n",
64
  "\n",
65
  "model_params = {\n",
@@ -70,6 +79,11 @@
70
  " 'MPT-30b': ('torch.bfloat16', 30.0),\n",
71
  " 'Phi-2': ('torch.float16', 2.78),\n",
72
  " 'Qwen-14b': ('torch.bfloat16', 14.167),\n",
 
 
 
 
 
73
  "}"
74
  ]
75
  },
 
46
  "new_df.model.unique()"
47
  ]
48
  },
49
+ {
50
+ "cell_type": "markdown",
51
+ "metadata": {},
52
+ "source": [
53
+ "- https://grabon.com/blog/claude-users-statistics/\n",
54
+ "- https://medium.com/@seanbetts/peering-inside-gpt-4-understanding-its-mixture-of-experts-moe-architecture-2a42eb8bdcb3\n",
55
+ "- https://www.cnbc.com/2023/05/16/googles-palm-2-uses-nearly-five-times-more-text-data-than-predecessor.html"
56
+ ]
57
+ },
58
  {
59
  "cell_type": "code",
60
  "execution_count": 4,
 
68
  " \"Phi-1.5\": \"microsoft/phi-1_5\",\n",
69
  " \"MPT-30b\": \"mosaicml/mpt-30b-instruct\",\n",
70
  " \"Phi-2\": \"microsoft/phi-2\",\n",
71
+ " \"Qwen-14b\": \"Qwen/Qwen-14B-Chat\",\n",
72
  "}\n",
73
  "\n",
74
  "model_params = {\n",
 
79
  " 'MPT-30b': ('torch.bfloat16', 30.0),\n",
80
  " 'Phi-2': ('torch.float16', 2.78),\n",
81
  " 'Qwen-14b': ('torch.bfloat16', 14.167),\n",
82
+ " 'Claude 2': ('?', 176),\n",
83
+ " 'Claude Instant': ('?', 60),\n",
84
+ " \"GPT 3.5 Turbo\": ('?', 175),\n",
85
+ " \"GPT 4 Turbo\": ('?', 1760),\n",
86
+ " 'PaLM 2': ('?', 340),\n",
87
  "}"
88
  ]
89
  },
lmsys/vicuna-13b-v1.3/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} RENAMED
File without changes
microsoft/phi-1_5/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} RENAMED
File without changes
microsoft/phi-2/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} RENAMED
File without changes
mistralai/Mistral-7B-Instruct-v0.1/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} RENAMED
File without changes
mosaicml/mpt-30b-instruct/{results_2024-01-19T15-58-52.json → results_2024-01-19T16-57-37.json} RENAMED
File without changes