Llama 2 70b

NLP Team Created at 03.02.2024 13:55

0.453

The overall result

The submission does not contain all the required tasks

Ratings for leaderboard tasks

The table will scroll to the left

Task name	Result	Metric
LCS	0.08	Accuracy
RCB	0.466 / 0.424	Accuracy F1 macro
USE	0.031	Grade norm
RWSD	0.5	Accuracy
PARus	0.744	Accuracy
ruTiE	0.453	Accuracy
MultiQ	0.185 / 0.041	F1 Exact match
CheGeKa	0.076 / 0	F1 Exact match
ruModAr	0.65	Exact match
ruMultiAr	0.216	Exact match
MathLogicQA	0.388	Accuracy
ruWorldTree	0.914 / 0.915	Accuracy F1 macro
ruOpenBookQA	0.818 / 0.817	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

0.495

Accuracy

ruMMLU

0.741

Accuracy

SimpleAr

0.965

Exact match

ruHumanEval

0.02 / 0.101 / 0.201

Pass@k

ruHHH

0.573

ruHateSpeech

0.585

ruDetox

0.341

ruEthics

	Correct	God	Ethical
Virtue	-0.113	-0.182	-0.143
Law	-0.124	-0.228	-0.171
Moral	-0.151	-0.21	-0.162
Justice	-0.065	-0.169	-0.145
Utilitarianism	-0.076	-0.153	-0.107

Information about the submission:

Mera version

Torch Version

The version of the codebase

CUDA version

Precision of the model weights

Seed

Butch

Transformers version

The number of GPUs and their type

Architecture

Team:

NLP Team

Name of the ML model:

Llama 2 70b

Link to the ML model:

https://huggingface.co/meta-llama/Llama-2-70b-hf

Additional links:

https://arxiv.org/abs/2307.09288

Architecture description:

Llama 2 is an auto-regressive language model that uses an optimized transformer architecture. Number of parameters 70b.

Description of the training:

Authors used custom training libraries, Meta's Research Super Cluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute. 1720320 GPU hours.

Pretrain data:

Llama 2 was pretrained on 2 trillion tokens of data from publicly available sources. Use standard transformer architecture, apply pre-normalization using RMSNorm, use the SwiGLU activation function, and rotary positional embeddings. The primary architectural differences from Llama 1 include increased context length and grouped-query attention (GQA).

Training Details:

Token counts refer to pretraining data only. All models are trained with a global batch-size of 4M tokens.

License:

A custom commercial license is available at: https://ai.meta.com/resources/models-and-libraries/llama-downloads/

Strategy, generation and parameters:

Code version v.1.1.0 All the parameters were not changed and are used as prepared by the organizers. Details: - 4 x NVIDIA A100 + accelerate - dtype float16 - Pytorch 2.0.1 + CUDA 11.7 - Transformers 4.36.2 - Context length 4096

Expand information

Ratings by subcategory

Metric: Accuracy

Model, team	Honest	Helpful	Harmless
Llama 2 70b NLP Team	0.557	0.508	0.655

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
Llama 2 70b NLP Team	0.6	0.75	0.7	0.714	0.762	0.9	0.867	0.765	0.9	1	0.727	0.6	0.4	0.462	0.545	0.7	0.8	0.778	0.8	1	0.4	0.865	0.8	0.745	0.7	0.818	0.625	0.714	1	0.636	0.5	0.778	0.8	0.8	0.818	1	0.7	0.81	0.4	0.6	0.823	0.9	0.8	0.6	0.938	0.6	1	0.5	0.6	0.9	0.636	1	0.853	0.867	0.5	0.485	0.778

Model, team	SIM	FL	STA
Llama 2 70b NLP Team	0.716	0.633	0.697

Coorect

Good

Ethical

Model, team	Virtue	Law	Moral	Justice	Utilitarianism
Llama 2 70b NLP Team	-0.113	-0.124	-0.151	-0.065	-0.076

Model, team	Virtue	Law	Moral	Justice	Utilitarianism
Llama 2 70b NLP Team	-0.182	-0.228	-0.21	-0.169	-0.153

Model, team	Virtue	Law	Moral	Justice	Utilitarianism
Llama 2 70b NLP Team	-0.143	-0.171	-0.162	-0.145	-0.107

Model, team	Women	Men	LGBT	Nationalities	Migrants	Other
Llama 2 70b NLP Team	0.583	0.571	0.706	0.595	0.429	0.574

Llama 2 70b

Ratings for leaderboard tasks

Evaluation on open tasks:

Information about the submission:

Team:

Name of the ML model:

Link to the ML model:

Additional links:

Architecture description:

Description of the training:

Pretrain data:

Training Details:

License:

Strategy, generation and parameters:

Ratings by subcategory

Confirm the deletion of the sub