adaptive-thinking
Adaptive_Skip_thinking_Reasoningannotations_creators:
found
language_creators:
llm-generated
languages:
ru
licenses:
unknown
multilinguality:
monolingual
pretty_name: Cerebras Adaptive Reasoning (Russian)
size_categories:
n-examples--1K
source_datasets: []
task_categories:
text-generation
reasoning
task_ids:
chain-of-thought
program-of-thought
skip-thinking
papers_with_code:
null
train_eval_split: []
configs:
default
Описание Датасета
cerebras-adaptive-reasoning-ru — это синтетический датасет, предназначенный для… See the full description on the dataset page: https://huggingface.co/datasets/Siesher/Adaptive_Skip_thinking_Reasoning.cortex-adaptive-thinking
Cortex Adaptive-Thinking, Interleaved Reasoning
A multi-domain instruction dataset that teaches a language model adaptive,
interleaved reasoning: instead of always emitting one long chain-of-thought up
front and then the answer, the model learns to place its thinking wherever a
real reasoner would, sometimes nowhere, sometimes once before answering,
sometimes mid-answer when a new sub-problem surfaces, sometimes as a closing
self-check, and sometimes across many think then act… See the full description on the dataset page: https://huggingface.co/datasets/PRATYUSH-BHARDWAJ/cortex-adaptive-thinking.