Meno-Lite-0.1

НГУ Created at 31.01.2026 03:53
0.555
The overall result
103
Place in the rating
Weak tasks:
168
RWSD
241
PARus
212
RCB
191
ruEthics
88
MultiQ
143
ruWorldTree
127
ruOpenBookQA
89
CheGeKa
191
ruMMLU
210
ruHateSpeech
118
ruDetox
180
ruHHH
117
ruTiE
157
ruHumanEval
137
USE
76
MathLogicQA
105
ruMultiAr
62
SimpleAr
56
LCS
182
BPS
213
ruModAr
145
MaMuRAMu
54
ruCodeEval
+19
Hide

Ratings for leaderboard tasks

The table will scroll to the left

Task name Result Metric
LCS 0.186 Accuracy
RCB 0.541 / 0.458 Accuracy F1 macro
USE 0.24 Grade norm
RWSD 0.569 Accuracy
PARus 0.818 Accuracy
ruTiE 0.794 Accuracy
MultiQ 0.536 / 0.403 F1 Exact match
CheGeKa 0.346 / 0.293 F1 Exact match
ruModAr 0.497 Exact match
MaMuRAMu 0.749 Accuracy
ruMultiAr 0.347 Exact match
ruCodeEval 0.377 / 0.569 / 0.622 Pass@k
MathLogicQA 0.666 Accuracy
ruWorldTree 0.949 / 0.76 Accuracy F1 macro
ruOpenBookQA 0.88 / 0.705 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 0.96 Accuracy
ruMMLU 0.624 Accuracy
SimpleAr 0.995 Exact match
ruHumanEval 0.109 / 0.193 / 0.22 Pass@k
ruHHH 0.742
ruHateSpeech 0.755
ruDetox 0.273
ruEthics
Correct God Ethical
Virtue 0.345 0.357 0.395
Law 0.346 0.327 0.391
Moral 0.368 0.353 0.415
Justice 0.338 0.313 0.375
Utilitarianism 0.277 0.294 0.344

Information about the submission

Mera version
v1.2.0
Torch Version
2.3.1
The version of the codebase
430295f
CUDA version
11.8
Precision of the model weights
bfloat16
Seed
1234
Batch
1
Transformers version
4.46.3
The number of GPUs and their type
1 x NVIDIA A100-PCIE-40GB
Architecture
hf

Team:

НГУ

Name of the ML model:

Meno-Lite-0.1

Model size

7.0B

Model type:

Opened

Architecture description:

Meno-Lite-0.1 is derived from RuadaptQwen2.5-7B-Lite-Beta and further adapted for Russian RAG pipelines, with a focus on information extraction, summarization, and question answering over complex legal and technical documents.

Description of the training:

Two-stage training: continued pretraining on RU/EN educational, legal, and sci-tech texts, then SFT on a custom instruction set targeting language skills (not world knowledge) to improve context use for document QA/RAG.

Pretrain data:

CPT data: sampled FineWeb-Edu (EN); RuLM (RU) subset selected for maximal FineWeb-Edu similarity using gte-multilingual-base; RU FinePDFs-edu; RuREBus (Dialogue’20) unlabeled text corpus. SFT data: NEREL converted to instructions + GPT-4o-mini synthetic entity normalization/definitions; LightRAG (GPT-4o) query logs on Habr + NSU site; plus MultiHopRAG and MTRAGEval train dialogs (EN).

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;

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.

Description of the template:
{%- if tools %} \n {{- '<|im_start|>system \n' }} \n {%- if messages[0]['role'] == 'system' %} \n {{- messages[0]['content'] }} \n {%- else %} \n {{- 'You are Meno, created by Ivan Bondarenko. You are a helpful assistant.' }} \n {%- endif %} \n {{- " \n \n# Tools \n \nYou may call one or more functions to assist with the user query. \n \nYou are provided with function signatures within <tools></tools> XML tags: \n<tools>" }} \n {%- for tool in tools %} \n {{- " \n" }} \n {{- tool | tojson }} \n {%- endfor %} \n {{- " \n</tools> \n \nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags: \n<tool_call> \n{"name": <function-name>, "arguments": <args-json-object>} \n</tool_call><|im_end|> \n" }} \n{%- else %} \n {%- if messages[0]['role'] == 'system' %} \n {{- '<|im_start|>system \n' + messages[0]['content'] + '<|im_end|> \n' }} \n {%- endif %} \n{%- endif %} \n{%- for message in messages %} \n {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %} \n {{- '<|im_start|>' + message.role + ' \n' + message.content + '<|im_end|>' + ' \n' }} \n {%- elif message.role == "assistant" %} \n {{- '<|im_start|>' + message.role }} \n {%- if message.content %} \n {{- ' \n' + message.content }} \n {%- endif %} \n {%- for tool_call in message.tool_calls %} \n {%- if tool_call.function is defined %} \n {%- set tool_call = tool_call.function %} \n {%- endif %} \n {{- ' \n<tool_call> \n{"name": "' }} \n {{- tool_call.name }} \n {{- '", "arguments": ' }} \n {{- tool_call.arguments | tojson }} \n {{- '} \n</tool_call>' }} \n {%- endfor %} \n {{- '<|im_end|> \n' }} \n {%- elif message.role == "tool" %} \n {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %} \n {{- '<|im_start|>user' }} \n {%- endif %} \n {{- ' \n<tool_response> \n' }} \n {{- message.content }} \n {{- ' \n</tool_response>' }} \n {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} \n {{- '<|im_end|> \n' }} \n {%- endif %} \n {%- endif %} \n{%- endfor %} \n{%- if add_generation_prompt %} \n {{- '<|im_start|>assistant \n' }} \n{%- endif %}

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
Meno-Lite-0.1
НГУ
0.5 0.467 0.667 0.233 0.033 0.367 0 - 0.033 0.033 0 0.033 0.133 0 0.1 0.267 0.033 0.067 0 0.033 0.033 0.733 0.367 0.1 0.033 0.55 0.067 0.2 0.333 0.4 0.433
Model, team Honest Helpful Harmless
Meno-Lite-0.1
НГУ
0.689 0.695 0.845
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
Meno-Lite-0.1
НГУ
0.548 0.488 0.776 0.833 0.716 0.761 0.777 0.627 0.722 0.632 0.509 0.444 0.4 0.63 0.771 0.662 0.72 0.715 0.467 0.672 0.332 0.807 0.46 0.653 0.491 0.73 0.422 0.602 0.645 0.45 0.83 0.752 0.693 0.808 0.691 0.688 0.49 0.826 0.53 0.606 0.788 0.581 0.634 0.69 0.803 0.602 0.765 0.478 0.426 0.618 0.55 0.802 0.682 0.723 0.81 0.727 0.746
Model, team SIM FL STA
Meno-Lite-0.1
НГУ
0.551 0.752 0.713
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
Meno-Lite-0.1
НГУ
0.622 0.842 0.617 0.667 0.75 0.655 0.724 0.754 0.712 0.723 0.731 0.683 0.492 0.783 0.678 0.716 0.692 0.733 0.614 0.772 0.491 0.881 0.756 0.793 0.711 0.758 0.795 0.614 0.93 0.822 0.844 0.808 0.714 0.895 0.697 0.804 0.889 0.733 0.684 0.708 0.824 0.857 0.8 0.844 0.879 0.889 0.897 0.818 0.846 0.877 0.778 0.797 0.797 0.74 0.674 0.649 0.822
Coorect
Good
Ethical
Model, team Virtue Law Moral Justice Utilitarianism
Meno-Lite-0.1
НГУ
0.345 0.346 0.368 0.338 0.277
Model, team Virtue Law Moral Justice Utilitarianism
Meno-Lite-0.1
НГУ
0.357 0.327 0.353 0.313 0.294
Model, team Virtue Law Moral Justice Utilitarianism
Meno-Lite-0.1
НГУ
0.395 0.391 0.415 0.375 0.344
Model, team Women Men LGBT Nationalities Migrants Other
Meno-Lite-0.1
НГУ
0.833 0.6 0.882 0.784 0.286 0.705