Datasets:

chris0809
/

scope-intent-routing-20k

Tasks:

Text Classification

Languages:

Chinese

English

Size:

10K<n<100K

License:

Dataset card Data Studio Files Files and versions

xet

Community

Dataset Preview

Duplicate

Split (1)

train

The full dataset viewer is not available (click to read why). Only showing a preview of the rows.

The dataset generation failed

Error code:   DatasetGenerationError
Exception:    CastError
Message:      Couldn't cast
created_at_utc: string
model: string
thinking_enabled: bool
generation_mode: string
rows: int64
reused_rows: int64
new_rows: int64
labels: struct<DIRECT_RESPONSE: int64, REQUIRES_SCOPE_CONTRACT: int64>
  child 0, DIRECT_RESPONSE: int64
  child 1, REQUIRES_SCOPE_CONTRACT: int64
splits: struct<train: int64, validation: int64, test: int64>
  child 0, train: int64
  child 1, validation: int64
  child 2, test: int64
languages: struct<zh: int64, en: int64>
  child 0, zh: int64
  child 1, en: int64
new_usage: struct<input_tokens: int64, output_tokens: int64, cache_read_tokens: int64>
  child 0, input_tokens: int64
  child 1, output_tokens: int64
  child 2, cache_read_tokens: int64
new_cost: struct<status: string, currency: string, known_total_cny: double, missing_requests: int64, priced_re (... 14 chars omitted)
  child 0, status: string
  child 1, currency: string
  child 2, known_total_cny: double
  child 3, missing_requests: int64
  child 4, priced_requests: int64
prior_4000_cost: struct<status: string, currency: string, known_total_cny: double, missing_requests: int64, priced_re (... 14 chars omitted)
  child 0, status: string
  child 1, currency: string
  child 2, known_total_cny: double
  child 3, missing_requests: int64
  child 4, priced_requests: int64
combined_known_cost_cny: double
elapsed_seconds: double
output: string
text: string
language: string
split: string
label: string
domain: string
id: string
label_id: int64
seed_id: string
difficulty: string
to
{'id': Value('string'), 'text': Value('string'), 'label': Value('string'), 'label_id': Value('int64'), 'split': Value('string'), 'seed_id': Value('string'), 'language': Value('string'), 'domain': Value('string'), 'difficulty': Value('string')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1827, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 613, in wrapped
                  for item in generator(*args, **kwargs):
                              ~~~~~~~~~^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              created_at_utc: string
              model: string
              thinking_enabled: bool
              generation_mode: string
              rows: int64
              reused_rows: int64
              new_rows: int64
              labels: struct<DIRECT_RESPONSE: int64, REQUIRES_SCOPE_CONTRACT: int64>
                child 0, DIRECT_RESPONSE: int64
                child 1, REQUIRES_SCOPE_CONTRACT: int64
              splits: struct<train: int64, validation: int64, test: int64>
                child 0, train: int64
                child 1, validation: int64
                child 2, test: int64
              languages: struct<zh: int64, en: int64>
                child 0, zh: int64
                child 1, en: int64
              new_usage: struct<input_tokens: int64, output_tokens: int64, cache_read_tokens: int64>
                child 0, input_tokens: int64
                child 1, output_tokens: int64
                child 2, cache_read_tokens: int64
              new_cost: struct<status: string, currency: string, known_total_cny: double, missing_requests: int64, priced_re (... 14 chars omitted)
                child 0, status: string
                child 1, currency: string
                child 2, known_total_cny: double
                child 3, missing_requests: int64
                child 4, priced_requests: int64
              prior_4000_cost: struct<status: string, currency: string, known_total_cny: double, missing_requests: int64, priced_re (... 14 chars omitted)
                child 0, status: string
                child 1, currency: string
                child 2, known_total_cny: double
                child 3, missing_requests: int64
                child 4, priced_requests: int64
              combined_known_cost_cny: double
              elapsed_seconds: double
              output: string
              text: string
              language: string
              split: string
              label: string
              domain: string
              id: string
              label_id: int64
              seed_id: string
              difficulty: string
              to
              {'id': Value('string'), 'text': Value('string'), 'label': Value('string'), 'label_id': Value('int64'), 'split': Value('string'), 'seed_id': Value('string'), 'language': Value('string'), 'domain': Value('string'), 'difficulty': Value('string')}
              because column names don't match
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1694, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1880, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

id string	text string	label string	split string	seed_id string	language string	domain string	difficulty string
d01-001	什么是关系型数据库？	DIRECT_RESPONSE	train	d01	zh	计算机科学	easy
d01-002	Can you explain what a NoSQL database is?	DIRECT_RESPONSE	train	d01	en	Technology	easy
d01-003	帮我定义一下图数据库。	DIRECT_RESPONSE	train	d01	zh	数据科学	medium
d01-004	What is an in-memory database?	DIRECT_RESPONSE	train	d01	en	Computing	medium
d01-005	请介绍一下分布式数据库的概念。	DIRECT_RESPONSE	train	d01	zh	IT架构	hard
d01-006	Explain the concept of a graph database to me.	DIRECT_RESPONSE	train	d01	en	Data Management	medium
d01-007	键值存储是什么？	DIRECT_RESPONSE	train	d01	zh	后端开发	easy
d01-008	Tell me about column-oriented databases.	DIRECT_RESPONSE	train	d01	en	Analytics	hard
d01-009	文档数据库和关系数据库有什么区别？	DIRECT_RESPONSE	train	d01	zh	软件开发	medium
d01-010	What constitutes a time-series database?	DIRECT_RESPONSE	train	d01	en	IoT	hard
d01-011	简单说说NewSQL数据库。	DIRECT_RESPONSE	train	d01	zh	数据库技术	hard
d01-012	Define OLAP databases for me.	DIRECT_RESPONSE	train	d01	en	Business Intelligence	medium
d01-013	什么是对象存储数据库？	DIRECT_RESPONSE	train	d01	zh	云存储	medium
d01-014	How does a document-oriented database work?	DIRECT_RESPONSE	train	d01	en	Software Engineering	medium
d01-015	解释一下内存数据库的原理。	DIRECT_RESPONSE	train	d01	zh	系统优化	hard
d01-016	What is a multi-model database?	DIRECT_RESPONSE	train	d01	en	Data Architecture	hard
d01-017	列式数据库适合什么场景？	DIRECT_RESPONSE	train	d01	zh	数据分析	medium
d01-018	Describe the architecture of a distributed SQL database.	DIRECT_RESPONSE	train	d01	en	Distributed Systems	hard
d01-019	什么是嵌入式数据库？	DIRECT_RESPONSE	train	d01	zh	物联网	easy
d01-020	What are the features of a key-value store?	DIRECT_RESPONSE	train	d01	en	Caching	easy
d01-021	帮我科普一下时序数据库。	DIRECT_RESPONSE	train	d01	zh	监控运维	medium
d01-022	Explain NewSQL to a beginner.	DIRECT_RESPONSE	train	d01	en	Database Theory	hard
d01-023	图数据库在处理社交网络时有什么优势？	DIRECT_RESPONSE	train	d01	zh	社交分析	medium
d01-024	What defines an object-oriented database?	DIRECT_RESPONSE	train	d01	en	Programming	hard
d01-025	什么是实时数据库？	DIRECT_RESPONSE	train	d01	zh	金融交易	medium
d01-026	Give me a definition of OLTP systems.	DIRECT_RESPONSE	train	d01	en	Enterprise IT	medium
d01-027	详细解释一下向量检索在数据库中的应用。	DIRECT_RESPONSE	train	d01	zh	AI应用	hard
d01-028	What is a cloud-native database?	DIRECT_RESPONSE	train	d01	en	Cloud Computing	medium
d01-029	多模数据库能同时处理哪些类型的数据？	DIRECT_RESPONSE	train	d01	zh	数据整合	hard
d01-030	Explain the concept of sharding in databases.	DIRECT_RESPONSE	train	d01	en	Scalability	hard
d01-031	什么是元组数据库？	DIRECT_RESPONSE	train	d01	zh	学术概念	hard
d01-032	What characterizes a hierarchical database?	DIRECT_RESPONSE	train	d01	en	Legacy Systems	medium
d01-033	解释一下HTAP数据库。	DIRECT_RESPONSE	train	d01	zh	混合事务处理	hard
d01-034	Describe a network model database.	DIRECT_RESPONSE	train	d01	en	Data Modeling	hard
d01-035	为什么选择内存数据库进行缓存？	DIRECT_RESPONSE	train	d01	zh	性能优化	medium
d01-036	What is a spatial database used for?	DIRECT_RESPONSE	train	d01	en	GIS	medium
d01-037	简单讲讲区块链数据库。	DIRECT_RESPONSE	train	d01	zh	Web3	medium
d01-038	Explain the role of a catalog database.	DIRECT_RESPONSE	train	d01	en	Information Retrieval	easy
d01-039	什么是宽列数据库？	DIRECT_RESPONSE	train	d01	zh	大数据	medium
d01-040	What are the pros and cons of NoSQL vs SQL?	DIRECT_RESPONSE	train	d01	en	Comparison	medium
d01-041	解释一下数据库中间件的作用。	DIRECT_RESPONSE	train	d01	zh	系统集成	medium
d01-042	Define a federated database system.	DIRECT_RESPONSE	train	d01	en	Data Integration	hard
d01-043	什么是基于文件的数据库？	DIRECT_RESPONSE	train	d01	zh	轻量级存储	easy
d01-044	What is a data lakehouse?	DIRECT_RESPONSE	train	d01	en	Modern Data Stack	medium
d01-045	聊聊持久化内存数据库。	DIRECT_RESPONSE	train	d01	zh	硬件加速	hard
d01-046	Explain the concept of eventual consistency.	DIRECT_RESPONSE	train	d01	en	Distributed Consensus	hard
d01-047	什么是只读数据库副本？	DIRECT_RESPONSE	train	d01	zh	高可用架构	easy
d01-048	What is a transactional memory database?	DIRECT_RESPONSE	train	d01	en	Concurrency Control	hard
d01-049	解释一下读写分离机制。	DIRECT_RESPONSE	train	d01	zh	负载均衡	medium
d01-050	Describe a replicated database.	DIRECT_RESPONSE	train	d01	en	Data Redundancy	easy
d01-051	什么是冷热数据分层存储？	DIRECT_RESPONSE	train	d01	zh	成本优化	medium
d01-052	What is a cold storage database solution?	DIRECT_RESPONSE	train	d01	en	Archive	medium
d01-053	讲解一下数据库连接池。	DIRECT_RESPONSE	train	d01	zh	资源管理	medium
d01-054	Explain the purpose of a connection pool.	DIRECT_RESPONSE	train	d01	en	Performance Tuning	medium
d01-055	什么是数据库索引？	DIRECT_RESPONSE	train	d01	zh	查询优化	easy
d01-056	What is a clustered index?	DIRECT_RESPONSE	train	d01	en	Index Structures	medium
d01-057	解释一下全文索引。	DIRECT_RESPONSE	train	d01	zh	搜索引擎	medium
d01-058	Describe full-text search capabilities.	DIRECT_RESPONSE	train	d01	en	Search Engines	medium
d01-059	什么是复合索引？	DIRECT_RESPONSE	train	d01	zh	数据库设计	medium
d01-060	What is a composite index?	DIRECT_RESPONSE	train	d01	en	Query Planning	medium
d01-061	解释一下覆盖索引。	DIRECT_RESPONSE	train	d01	zh	执行计划	hard
d01-062	Define a covering index.	DIRECT_RESPONSE	train	d01	en	Optimization	hard
d01-063	什么是B+树索引？	DIRECT_RESPONSE	train	d01	zh	数据结构	medium
d01-064	What is a B+ tree index?	DIRECT_RESPONSE	train	d01	en	Data Structures	medium
d01-065	解释一下哈希索引。	DIRECT_RESPONSE	train	d01	zh	快速查找	medium
d01-066	Describe a hash index.	DIRECT_RESPONSE	train	d01	en	Lookup Methods	medium
d01-067	什么是位图索引？	DIRECT_RESPONSE	train	d01	zh	数据仓库	hard
d01-068	What is a bitmap index?	DIRECT_RESPONSE	train	d01	en	Warehousing	hard
d01-069	解释一下空间索引。	DIRECT_RESPONSE	train	d01	zh	地理信息	hard
d01-070	Explain spatial indexing techniques.	DIRECT_RESPONSE	train	d01	en	Spatial Data	hard
d01-071	什么是外键约束？	DIRECT_RESPONSE	train	d01	zh	数据完整性	easy
d01-072	What is a foreign key constraint?	DIRECT_RESPONSE	train	d01	en	Relational Integrity	easy
d01-073	解释一下主键自增。	DIRECT_RESPONSE	train	d01	zh	标识符管理	easy
d01-074	Describe auto-increment primary keys.	DIRECT_RESPONSE	train	d01	en	Key Generation	easy
d01-075	什么是唯一约束？	DIRECT_RESPONSE	train	d01	zh	数据清洗	easy
d01-076	What is a unique constraint?	DIRECT_RESPONSE	train	d01	en	Data Quality	easy
d01-077	解释一下非空约束。	DIRECT_RESPONSE	train	d01	zh	Schema Design	easy
d01-078	Define a not null constraint.	DIRECT_RESPONSE	train	d01	en	Validation	easy
d01-079	什么是检查约束？	DIRECT_RESPONSE	train	d01	zh	业务规则	medium
d01-080	What is a check constraint?	DIRECT_RESPONSE	train	d01	en	Business Logic	medium
d01-081	解释一下触发器。	DIRECT_RESPONSE	train	d01	zh	自动化逻辑	medium
d01-082	Explain database triggers.	DIRECT_RESPONSE	train	d01	en	Automation	medium
d01-083	什么是存储过程？	DIRECT_RESPONSE	train	d01	zh	服务端逻辑	medium
d01-084	What is a stored procedure?	DIRECT_RESPONSE	train	d01	en	Server-Side Code	medium
d01-085	解释一下视图。	DIRECT_RESPONSE	train	d01	zh	虚拟表	easy
d01-086	Describe a database view.	DIRECT_RESPONSE	train	d01	en	Abstraction	easy
d01-087	什么是物化视图？	DIRECT_RESPONSE	train	d01	zh	性能预计算	medium
d01-088	What is a materialized view?	DIRECT_RESPONSE	train	d01	en	Pre-computation	medium
d01-089	解释一下事务隔离级别。	DIRECT_RESPONSE	train	d01	zh	并发控制	hard
d01-090	Explain transaction isolation levels.	DIRECT_RESPONSE	train	d01	en	Concurrency	hard
d01-091	什么是脏读？	DIRECT_RESPONSE	train	d01	zh	一致性	medium
d01-092	What is a dirty read?	DIRECT_RESPONSE	train	d01	en	Anomalies	medium
d01-093	解释一下幻读。	DIRECT_RESPONSE	train	d01	zh	数据一致性	hard
d01-094	Describe phantom reads.	DIRECT_RESPONSE	train	d01	en	Isolation Issues	hard
d01-095	向量数据库是啥？	DIRECT_RESPONSE	train	d01	zh	技术科普	easy
d01-096	说说向量数据库咋回事。	DIRECT_RESPONSE	train	d01	zh	技术科普	easy
d01-097	给咱讲讲向量数据库呗。	DIRECT_RESPONSE	train	d01	zh	技术科普	easy
d01-098	Vector DB, what's that?	DIRECT_RESPONSE	train	d01	en	tech_support	easy
d01-099	Can you explain vector databases simply?	DIRECT_RESPONSE	train	d01	en	tech_support	easy
d01-100	What on earth is a vector database?	DIRECT_RESPONSE	train	d01	en	tech_support	medium

End of preview.

Scope Intent Routing 20K

This bilingual synthetic dataset trains a high-recall router that decides whether a user request needs structured scope resolution before execution.

DIRECT_RESPONSE (label_id=0): the request can be answered from the text already supplied by the user and needs no external state.
REQUIRES_SCOPE_CONTRACT (label_id=1): the request needs tools, files, accounts, web access, databases, code execution, or other external reads or writes, so the system should first resolve objects, conditions, and operations.

The dataset contains 20,000 unique examples, balanced between the two labels. Its train/validation/test splits contain 14,000/3,000/3,000 rows and use disjoint seed families.

Labels were fixed by 80 human-authored seed families. Qwen3.7-Flash, with thinking disabled, generated paraphrases inside each fixed family and did not choose the label. The dataset is intended for conservative routing, where missing a request that needs scope resolution is more costly than occasionally invoking the resolver unnecessarily.

This is synthetic data and does not represent production traffic. Reported model metrics should be described as held-out synthetic-family results, not real-world accuracy.

Downloads last month: 24

Total file size:

5.12 MB

Models trained or fine-tuned on chris0809/scope-intent-routing-20k

chris0809/scope-intent-distilmbert

Text Classification • 0.1B • Updated 2 days ago • 15