patrickvonplaten commited on
Commit
c76e8f8
·
verified ·
1 Parent(s): aedf0ca

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +54 -1
README.md CHANGED
@@ -34,6 +34,59 @@ Learn more about Mistral Small in our [blog post](https://mistral.ai/news/mistra
34
  - **Context Window:** A 32k context window.
35
  - **Tokenizer:** Utilizes a Tekken tokenizer with a 131k vocabulary size.
36
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
37
  ## The Mistral AI Team
38
 
39
- Albert Jiang, Alexandre Sablayrolles, Alexis Tacnet, Alok Kothari, Antoine Roux, Arthur Mensch, Audrey Herblin-Stoop, Augustin Garreau, Austin Birky, Bam4d, Baptiste Bout, Baudouin de Monicault, Blanche Savary, Carole Rambaud, Caroline Feldman, Devendra Singh Chaplot, Diego de las Casas, Diogo Costa, Eleonore Arcelin, Emma Bou Hanna, Etienne Metzger, Gaspard Blanchet, Gianna Lengyel, Guillaume Bour, Guillaume Lample, Harizo Rajaona, Henri Roussez, Hichem Sattouf, Ian Mack, Jean-Malo Delignon, Jessica Chudnovsky, Justus Murke, Kartik Khandelwal, Lawrence Stewart, Louis Martin, Louis Ternon, Lucile Saulnier, Lélio Renard Lavaud, Margaret Jennings, Marie Pellat, Marie Torelli, Marie-Anne Lachaux, Marjorie Janiewicz, Mickaël Seznec, Nicolas Schuhl, Niklas Muhs, Olivier de Garrigues, Patrick von Platen, Paul Jacob, Pauline Buche, Pavan Kumar Reddy, Perry Savas, Pierre Stock, Romain Sauvestre, Sagar Vaze, Sandeep Subramanian, Saurabh Garg, Sophia Yang, Szymon Antoniak, Teven Le Scao, Thibault Schueller, Thibaut Lavril, Thomas Wang, Théophile Gervet, Timothée Lacroix, Valera Nemychnikova, Wendy Shang, William El Sayed, William Marshall
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
34
  - **Context Window:** A 32k context window.
35
  - **Tokenizer:** Utilizes a Tekken tokenizer with a 131k vocabulary size.
36
 
37
+ ## Benchmark Results
38
+
39
+ | Benchmark | Metric | Mistral-Small-24B-Base |
40
+ | ------------------------------ | ------------- | ----------- |
41
+ | [MMLU][mmlu] | 5-shot | 80.73 |
42
+ | [MMLU Pro][mmlu_pro] | 5-shot, CoT | 54.37 |
43
+ | [GPQA Main][gpqa] | 5-shot, CoT | 34.37 |
44
+ | [TriviaQA][triviaqa] | 5-shot | 80.32 |
45
+ | [ARC-c][arc] | 0-shot | 91.29 |
46
+ | [TriviaQA][triviaqa] | 5-shot | 76.6 |
47
+ | [MBPP][mbpp] | pass@1 | 69.64 |
48
+ | [GSM8K][gsm8k] | 5-shot, maj@1 | 80.73 |
49
+ | [MATH][math] | 4-shot, MaJ | 45.98 |
50
+ | [AGIEval][agieval] | - | 65.80 |
51
+ | ------------------------------ | ------------- | ----------- |
52
+ | French MMLU | - | 78.03 |
53
+ | German MMLU | - | 77.69 |
54
+ | Spanish MMLU | - | 78.86 |
55
+ | Russian MMLU | - | 75.64 |
56
+ | Chinese MMLU | - | 70.35 |
57
+ | Korean MMLU | - | 56.42 |
58
+ | Japanese MMLU | - | 74.46 |
59
+
60
+
61
  ## The Mistral AI Team
62
 
63
+ Albert Jiang, Alexandre Sablayrolles, Alexis Tacnet, Alok Kothari, Antoine Roux, Arthur Mensch, Audrey Herblin-Stoop, Augustin Garreau, Austin Birky, Bam4d, Baptiste Bout, Baudouin de Monicault, Blanche Savary, Carole Rambaud, Caroline Feldman, Devendra Singh Chaplot, Diego de las Casas, Diogo Costa, Eleonore Arcelin, Emma Bou Hanna, Etienne Metzger, Gaspard Blanchet, Gianna Lengyel, Guillaume Bour, Guillaume Lample, Harizo Rajaona, Henri Roussez, Hichem Sattouf, Ian Mack, Jean-Malo Delignon, Jessica Chudnovsky, Justus Murke, Kartik Khandelwal, Lawrence Stewart, Louis Martin, Louis Ternon, Lucile Saulnier, Lélio Renard Lavaud, Margaret Jennings, Marie Pellat, Marie Torelli, Marie-Anne Lachaux, Marjorie Janiewicz, Mickaël Seznec, Nicolas Schuhl, Niklas Muhs, Olivier de Garrigues, Patrick von Platen, Paul Jacob, Pauline Buche, Pavan Kumar Reddy, Perry Savas, Pierre Stock, Romain Sauvestre, Sagar Vaze, Sandeep Subramanian, Saurabh Garg, Sophia Yang, Szymon Antoniak, Teven Le Scao, Thibault Schueller, Thibaut Lavril, Thomas Wang, Théophile Gervet, Timothée Lacroix, Valera Nemychnikova, Wendy Shang, William El Sayed, William Marshall
64
+
65
+
66
+ [mmlu]: https://arxiv.org/abs/2009.03300
67
+ [hellaswag]: https://arxiv.org/abs/1905.07830
68
+ [piqa]: https://arxiv.org/abs/1911.11641
69
+ [socialiqa]: https://arxiv.org/abs/1904.09728
70
+ [boolq]: https://arxiv.org/abs/1905.10044
71
+ [winogrande]: https://arxiv.org/abs/1907.10641
72
+ [commonsenseqa]: https://arxiv.org/abs/1811.00937
73
+ [openbookqa]: https://arxiv.org/abs/1809.02789
74
+ [arc]: https://arxiv.org/abs/1911.01547
75
+ [triviaqa]: https://arxiv.org/abs/1705.03551
76
+ [naturalq]: https://github.com/google-research-datasets/natural-questions
77
+ [humaneval]: https://arxiv.org/abs/2107.03374
78
+ [mbpp]: https://arxiv.org/abs/2108.07732
79
+ [gsm8k]: https://arxiv.org/abs/2110.14168
80
+ [realtox]: https://arxiv.org/abs/2009.11462
81
+ [bold]: https://arxiv.org/abs/2101.11718
82
+ [crows]: https://aclanthology.org/2020.emnlp-main.154/
83
+ [bbq]: https://arxiv.org/abs/2110.08193v2
84
+ [winogender]: https://arxiv.org/abs/1804.09301
85
+ [truthfulqa]: https://arxiv.org/abs/2109.07958
86
+ [winobias]: https://arxiv.org/abs/1804.06876
87
+ [math]: https://arxiv.org/abs/2103.03874
88
+ [agieval]: https://arxiv.org/abs/2304.06364
89
+ [big-bench]: https://arxiv.org/abs/2206.04615
90
+ [toxigen]: https://arxiv.org/abs/2203.09509
91
+ [mmlu_pro]: https://arxiv.org/abs/2406.01574
92
+ [gpqa]: https://arxiv.org/abs/2311.12022