Yi-Coder-9B-Chat

MERA Created at 04.11.2024 18:05
0.398
The overall result
328
Place in the rating
Weak tasks:
82
RWSD
399
PARus
338
RCB
386
ruEthics
362
MultiQ
455
ruWorldTree
451
ruOpenBookQA
418
CheGeKa
470
ruMMLU
360
ruHateSpeech
352
ruDetox
331
ruHHH
388
ruTiE
43
ruHumanEval
312
USE
329
MathLogicQA
372
ruMultiAr
415
SimpleAr
83
LCS
86
BPS
364
ruModAr
445
MaMuRAMu
52
ruCodeEval
+19
Hide

Ratings for leaderboard tasks

The table will scroll to the left

Task name Result Metric
LCS 0.164 Accuracy
RCB 0.505 / 0.443 Accuracy F1 macro
USE 0.123 Grade norm
RWSD 0.615 Accuracy
PARus 0.642 Accuracy
ruTiE 0.558 Accuracy
MultiQ 0.3 / 0.159 F1 Exact match
CheGeKa 0.029 / 0.012 F1 Exact match
ruModAr 0.396 Exact match
MaMuRAMu 0.457 Accuracy
ruMultiAr 0.188 Exact match
ruCodeEval 0.384 / 0.658 / 0.713 Pass@k
MathLogicQA 0.373 Accuracy
ruWorldTree 0.619 / 0.61 Accuracy F1 macro
ruOpenBookQA 0.543 / 0.533 Accuracy F1 macro

Evaluation on open tasks:

Go to the ratings by subcategory

The table will scroll to the left

Task name Result Metric
BPS 0.984 Accuracy
ruMMLU 0.377 Accuracy
SimpleAr 0.82 Exact match
ruHumanEval 0.414 / 0.696 / 0.756 Pass@k
ruHHH 0.607
ruHateSpeech 0.623
ruDetox 0.136
ruEthics
Correct God Ethical
Virtue 0.125 0.213 0.163
Law 0.118 0.19 0.165
Moral 0.129 0.214 0.163
Justice 0.128 0.192 0.131
Utilitarianism 0.114 0.152 0.144

Information about the submission

Mera version
v.1.2.0
Torch Version
2.4.0
The version of the codebase
9b26db97
CUDA version
12.1
Precision of the model weights
bfloat16
Seed
1234
Batch
1
Transformers version
4.44.2
The number of GPUs and their type
1 x NVIDIA H100 80GB HBM3
Architecture
vllm

Team:

MERA

Name of the ML model:

Yi-Coder-9B-Chat

Model size

9.0B

Model type:

Opened

SFT

Additional links:

https://01-ai.github.io/blog.html?post=en/2024-09-05-A-Small-but-Mighty-LLM-for-Code.md

Architecture description:

Yi-Coder-9B builds upon Yi-9B with an additional 2.4T high-quality tokens, meticulously sourced from a repository-level code corpus on GitHub and code-related data filtered from CommonCrawl.

Description of the training:

-

Pretrain data:

Continue pretrained on 2.4 Trillion high-quality tokens over 52 major programming languages.

License:

apache-2.0

Inference parameters

Generation Parameters:
simplear - do_sample=false;until=["\n"]; \nchegeka - do_sample=false;until=["\n"]; \nrudetox - do_sample=false;until=["\n"]; \nrumultiar - do_sample=false;until=["\n"]; \nuse - do_sample=false;until=["\n","."]; \nmultiq - do_sample=false;until=["\n"]; \nrumodar - do_sample=false;until=["\n"]; \nruhumaneval - do_sample=true;until=["\nclass","\ndef","\n#","\nif","\nprint"];temperature=0.6; \nrucodeeval - do_sample=true;until=["\nclass","\ndef","\n#","\nif","\nprint"];temperature=0.6;

The size of the context:
simplear, bps, lcs, chegeka, mathlogicqa, parus, rcb, rudetox, ruhatespeech, rummlu, ruworldtree, ruopenbookqa, rumultiar, use, rwsd, mamuramu, multiq, rumodar, ruethics, ruhhh, ruhumaneval, rucodeeval, rutie - 131072

System prompt:
Реши задачу по инструкции ниже. Не давай никаких объяснений и пояснений к своему ответу. Не пиши ничего лишнего. Пиши только то, что указано в инструкции. Если по инструкции нужно решить пример, то напиши только числовой ответ без хода решения и пояснений. Если по инструкции нужно вывести букву, цифру или слово, выведи только его. Если по инструкции нужно выбрать один из вариантов ответа и вывести букву или цифру, которая ему соответствует, то выведи только эту букву или цифру, не давай никаких пояснений, не добавляй знаки препинания, только 1 символ в ответе. Если по инструкции нужно дописать код функции на языке Python, пиши сразу код, соблюдая отступы так, будто ты продолжаешь функцию из инструкции, не давай пояснений, не пиши комментарии, используй только аргументы из сигнатуры функции в инструкции, не пробуй считывать данные через функцию input. Не извиняйся, не строй диалог. Выдавай только ответ и ничего больше.

Ratings by subcategory

Metric: Grade Norm
Model, team 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 8_0 8_1 8_2 8_3 8_4
Yi-Coder-9B-Chat
MERA
0.333 0.133 0.567 0.1 0 0.033 0 - 0 0 0.067 0 0.133 0 0.1 0.267 0.033 0 0.067 0.033 0.033 0.1 0 0 0.033 0.208 0.2 0.133 0.333 0.133 0.233
Model, team Honest Helpful Harmless
Yi-Coder-9B-Chat
MERA
0.557 0.712 0.552
Model, team Anatomy Virology Astronomy Marketing Nutrition Sociology Management Philosophy Prehistory Human aging Econometrics Formal logic Global facts Jurisprudence Miscellaneous Moral disputes Business ethics Biology (college) Physics (college) Human Sexuality Moral scenarios World religions Abstract algebra Medicine (college) Machine learning Medical genetics Professional law PR Security studies Chemistry (школьная) Computer security International law Logical fallacies Politics Clinical knowledge Conceptual_physics Math (college) Biology (high school) Physics (high school) Chemistry (high school) Geography (high school) Professional medicine Electrical engineering Elementary mathematics Psychology (high school) Statistics (high school) History (high school) Math (high school) Professional accounting Professional psychology Computer science (college) World history (high school) Macroeconomics Microeconomics Computer science (high school) European history Government and politics
Yi-Coder-9B-Chat
MERA
0.281 0.301 0.447 0.556 0.402 0.418 0.544 0.421 0.401 0.359 0.377 0.381 0.29 0.472 0.442 0.413 0.39 0.326 0.233 0.313 0.246 0.327 0.28 0.353 0.321 0.42 0.296 0.407 0.433 0.26 0.55 0.529 0.387 0.424 0.408 0.402 0.32 0.423 0.358 0.31 0.424 0.29 0.524 0.435 0.431 0.361 0.475 0.326 0.33 0.361 0.39 0.481 0.359 0.37 0.68 0.424 0.399
Model, team SIM FL STA
Yi-Coder-9B-Chat
MERA
0.713 0.49 0.447
Model, team Anatomy Virology Astronomy Marketing Nutrition Sociology Managment Philosophy Pre-History Gerontology Econometrics Formal logic Global facts Jurisprudence Miscellaneous Moral disputes Business ethics Bilology (college) Physics (college) Human sexuality Moral scenarios World religions Abstract algebra Medicine (college) Machine Learning Genetics Professional law PR Security Chemistry (college) Computer security International law Logical fallacies Politics Clinical knowledge Conceptual physics Math (college) Biology (high school) Physics (high school) Chemistry (high school) Geography (high school) Professional medicine Electrical Engineering Elementary mathematics Psychology (high school) Statistics (high school) History (high school) Math (high school) Professional Accounting Professional psychology Computer science (college) World history (high school) Macroeconomics Microeconomics Computer science (high school) Europe History Government and politics
Yi-Coder-9B-Chat
MERA
0.356 0.436 0.383 0.546 0.421 0.397 0.483 0.386 0.365 0.415 0.526 0.567 0.342 0.465 0.345 0.346 0.495 0.333 0.263 0.526 0.228 0.492 0.444 0.391 0.533 0.379 0.551 0.491 0.684 0.467 0.622 0.372 0.393 0.456 0.348 0.518 0.6 0.6 0.421 0.462 0.461 0.286 0.533 0.622 0.776 0.711 0.448 0.636 0.631 0.614 0.756 0.29 0.582 0.545 0.395 0.304 0.467
Coorect
Good
Ethical
Model, team Virtue Law Moral Justice Utilitarianism
Yi-Coder-9B-Chat
MERA
0.125 0.118 0.129 0.128 0.114
Model, team Virtue Law Moral Justice Utilitarianism
Yi-Coder-9B-Chat
MERA
0.213 0.19 0.214 0.192 0.152
Model, team Virtue Law Moral Justice Utilitarianism
Yi-Coder-9B-Chat
MERA
0.163 0.165 0.163 0.131 0.144
Model, team Women Men LGBT Nationalities Migrants Other
Yi-Coder-9B-Chat
MERA
0.602 0.714 0.647 0.703 0.429 0.574