Datasets:

Perfectyash
/

human-style-cyber-incident-judgment

incident_id int64 1 50	system_type stringlengths 7 21	signal_strength stringclasses 3 values	anomaly_type stringlengths 7 21	human_pressure_level stringclasses 3 values	past_similar_cases stringclasses 2 values	initial_human_thought stringlengths 10 25	reasoning_summary stringlengths 15 44	final_decision stringlengths 6 21	decision_confidence float64 0.5 0.99
1	Corporate Email	Medium	Suspicious attachment	Low	Yes	Could be phishing	Attachment resembles known phishing patterns	Quarantine email	0.91
2	Cloud Server	Strong	Unauthorized access	High	Yes	Critical breach	Login from unknown region with admin account	Isolate server	0.96
3	Mobile App	Weak	Login delay	Low	No	Probably network issue	No repeated malicious indicators	Monitor only	0.62
4	Payment Gateway	Strong	Transaction anomaly	High	Yes	Fraud likely	Multiple high-value transactions in seconds	Block transactions	0.94
5	Healthcare System	Medium	Data access spike	Medium	Yes	Possible misuse	Non-duty hour access to patient data	Audit user activity	0.89
6	University Network	Weak	Port scan	Low	No	Student experiment	Scan limited and non-aggressive	Log event	0.58
7	E-commerce Platform	Strong	Bot traffic	Medium	Yes	Automated attack	Pattern matches known bots	Enable bot mitigation	0.93
8	Corporate VPN	Medium	Concurrent login	High	Yes	Credential leak suspected	Impossible travel scenario	Force password reset	0.92
9	IoT Network	Weak	Device disconnect	Low	No	Hardware issue	No threat indicators observed	Ignore	0.55
10	Social Media Platform	Strong	Account takeover	High	Yes	Confirmed breach	Password changed from new IP	Lock account	0.95
11	Financial System	Medium	Policy violation	Medium	Yes	Insider risk	Repeated access to restricted files	Notify compliance	0.88
12	Cloud Storage	Strong	Mass download	High	Yes	Data exfiltration	Large data pulled rapidly	Revoke access	0.97
13	HR System	Weak	Login retry	Low	No	User mistake	Few retries without pattern	No action	0.52
14	Web Server	Medium	SQL injection attempt	Medium	Yes	Attack probe detected	Suspicious query patterns	Block IP	0.9
15	Industrial Control	Strong	Command override	High	Yes	Severe risk	Unauthorized command execution	Shutdown system	0.98
16	Corporate Chat	Weak	Unknown link	Low	No	Probably safe	No malicious reputation	Warn user	0.6
17	Retail POS	Medium	Card mismatch	Medium	Yes	Fraud suspicion	Location mismatch detected	Flag transaction	0.87
18	DevOps Pipeline	Strong	Code tampering	High	Yes	Supply chain risk	Unexpected commit detected	Rollback deployment	0.96
19	School Portal	Weak	Multiple logins	Low	No	Shared device issue	Normal school hours	Monitor	0.59
20	Email Gateway	Strong	Malware signature	High	Yes	Confirmed malware	Hash matches malware DB	Delete email	0.99
21	CRM System	Medium	API abuse	Medium	Yes	Automation misuse	Excessive API calls	Throttle access	0.85
22	Streaming Platform	Weak	Buffer anomaly	Low	No	Performance issue	No security flags	Ignore	0.5
23	Smart Home Hub	Medium	Unauthorized pairing	Medium	No	Potential intrusion	Unknown device detected	Block device	0.86
24	Corporate WiFi	Strong	Rogue AP	High	Yes	Active attack	Fake AP detected	Disable access point	0.95
25	Log Server	Weak	Log gap	Low	No	System delay	No correlated events	No action	0.54
26	Banking App	Strong	Rooted device login	High	Yes	High fraud risk	Root access detected	Deny login	0.94
27	Video Conference	Medium	Session hijack	Medium	Yes	Privacy risk	Unexpected participant join	Terminate session	0.89
28	Inventory System	Weak	Stock mismatch	Low	No	Sync error suspected	No malicious behavior	Investigate later	0.57
29	Travel Portal	Strong	Location spoofing	High	Yes	Fraud attempt	VPN abuse detected	Block account	0.93
30	Ad Platform	Medium	Click fraud	Medium	Yes	Suspicious behavior	Automated clicking pattern	Blacklist IPs	0.88
31	Smart Meter	Weak	Data delay	Low	No	Connectivity issue	No intrusion signals	Ignore	0.51
32	Government Portal	Strong	Identity spoofing	High	Yes	Serious breach	Fake documents uploaded	Freeze application	0.97
33	Forum Website	Weak	Spam post	Low	No	Normal spam	Basic spam behavior	Delete post	0.56
34	AI Model API	Medium	Prompt abuse	Medium	Yes	Policy evasion attempt	Repeated boundary probing	Rate limit user	0.84
35	Logistics System	Strong	Route manipulation	High	Yes	Operational sabotage	Unauthorized changes detected	Rollback routes	0.96
36	Learning Platform	Weak	Quiz cheating	Low	No	Student behavior	Minor violation	Warn student	0.53
37	Stock Trading App	Strong	Market manipulation	High	Yes	Illegal activity	Abnormal order patterns	Freeze trades	0.98
38	Messaging App	Medium	Mass invites	Medium	Yes	Spam campaign	Invite velocity abnormal	Restrict account	0.87
39	Payroll System	Weak	Salary mismatch	Low	No	Input error	No malicious intent	Manual correction	0.55
40	News Website	Medium	Traffic surge	Medium	Yes	Possible DDoS	Traffic pattern unusual	Enable rate limiting	0.86
41	IoT Camera	Strong	Remote access	High	Yes	Privacy breach	Unknown external access	Disable camera	0.95
42	Cloud Console	Medium	Permission escalation	Medium	Yes	Privilege abuse	Risk escalation observed	Revoke permissions	0.9
43	Forum Admin Panel	Weak	Failed login	Low	No	Typing error	Single attempt only	No action	0.5
44	Search Engine Ads	Medium	Ad fraud	Medium	Yes	Suspicious spend	CTR anomaly detected	Suspend campaign	0.88
45	Airport System	Strong	Security override	High	Yes	Critical threat	Unauthorized override detected	Emergency shutdown	0.99
46	Chatbot System	Weak	Odd response	Low	No	Model glitch suspected	No abuse evidence	Monitor	0.52
47	CRM API	Medium	Token misuse	Medium	Yes	Access risk	Token shared externally	Revoke token	0.89
48	Cloud Backup	Strong	Deletion attempt	High	Yes	Ransomware signal	Mass delete initiated	Lock backups	0.97
49	Game Server	Weak	Latency spike	Low	No	Load issue	No malicious indicators	Ignore	0.51
50	Enterprise IAM	Strong	Privilege escalation	High	Yes	Major breach risk	Unauthorized role assignment	Revert and alert SOC	0.98

YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/datasets-cards)

Human-Style Cyber Incident Judgment Dataset

📌 Overview

This dataset represents human-like reasoning in cybersecurity incident response.
Each entry simulates how a security analyst thinks, reasons, and decides under different levels of pressure and risk.

Unlike traditional datasets that focus on logs or alerts, this dataset focuses on:

intuition
contextual reasoning
experience-based judgment
confidence in decision-making

🧠 Uniqueness

Fully synthetic and original
Not copied or derived from real logs
Designed manually for AI reasoning research
Focuses on human decision process, not automation

📊 Dataset Structure

Each row represents a single simulated cyber incident.

Columns

incident_id – Unique identifier
system_type – Affected system or platform
signal_strength – Weak / Medium / Strong
anomaly_type – Type of suspicious activity
human_pressure_level – Analyst stress level
past_similar_cases – Prior experience indicator
initial_human_thought – First intuitive reaction
reasoning_summary – Explanation behind the decision
final_decision – Action taken
decision_confidence – Confidence score (0–1)

🎯 Intended Use

Explainable AI (XAI)
Cybersecurity research
SOC analyst simulation
AI/ML reasoning models
Educational and academic projects

📜 License

This dataset is released under CC0 1.0 (Public Domain).

You are free to:

Use
Modify
Distribute
Build commercial or non-commercial projects

No attribution is required.

⚠️ Disclaimer

This dataset is synthetic and does not contain real user data, real systems, or sensitive information.
Any resemblance to real incidents is purely coincidental.

Downloads last month: 24

Size of downloaded dataset files:

5.69 kB

Size of the auto-converted Parquet files:

9.84 kB

Number of rows: