The technique that enables LLMs to generate code by learning from programming language corpora is called Code LLM/Generation.
- True
- False
Answer: True
Code LLMs enable natural language to code generation, debugging, documentation.
GK Questions
Filter by category, type, and difficulty. Reading is open for everyone.
Category
Difficulty
Type
Per page
Instruction tuning adapts LLMs to follow user instructions by training on prompt-________ pairs.
- Response
- Query
- Input
- Output
Answer: Response
Instruction tuning fine-tunes LLMs on instruction datasets, improving task completion and helpfulness.
The international organization that develops global internet governance policies is called Internet Corporation for Assigned Names and Numbers.
- True
- False
Answer: True
ICANN coordinates domain names, IP addresses, protocol parameters. Multi-stakeholder model.
The Digital India Act proposal includes provisions for regulating algorithmic transparency and ________.
- Accountability
- Navigation Satellite
- RTI Online
- Knowledge
Answer: Accountability
Digital India Act aims to address platform accountability, user rights, emerging tech governance.
The Indian initiative for promoting electric vehicle adoption through subsidies is called FAME India.
- True
- False
Answer: True
FAME provides demand incentives for EVs and charging infrastructure.
The process of documenting and analyzing security incidents for improvement is called Post-Incident Review.
- True
- False
Answer: True
Post-incident review captures lessons learned: what happened, response effectiveness, root causes, improvements. Critical for continuous security program maturity.
The technique that enables LLMs to generate multiple plausible responses and select the best is called Beam/Nucleus Sampling.
- True
- False
Answer: True
Advanced decoding: beam search explores multiple hypotheses, nucleus sampling truncates low-probability tokens. Balance diversity and coherence.
The initiative for global cooperation on cybersecurity capacity building is called Global Cybersecurity Index.
- True
- False
Answer: True
ITU Global Cybersecurity Index assesses national cybersecurity commitment across legal, technical, organizational, capacity building, cooperation dimensions.
The practice of simulating cyberattacks to test organizational defenses is called Red Team Testing.
- True
- False
Answer: True
Red teaming conducts adversarial simulations (cyber, physical, social) to identify gaps in detection, response, resilience. Distinct from vulnerability scanning.
Retail CBDC (e₹-R) enables programmable welfare payments with usage restrictions to prevent misuse.
- True
- False
Answer: True
Programmable CBDC allows conditional transfers for specific goods/services (food, education), enhancing policy effectiveness while requiring careful privacy design.
Which ISRO center in Sriharikota handles all satellite and human spaceflight launches?
- VSSC
- URSC
- SDSC-SHAR
- LPSC
Answer: SDSC-SHAR
Satish Dhawan Space Centre (SDSC-SHAR) in Andhra Pradesh is ISRO's primary launch port. Handles PSLV, GSLV, SSLV, and future human spaceflight launches.
In-situ resource utilization (ISRU) enables production of rocket propellant from lunar water ice.
- True
- False
Answer: True
ISRU extracts water ice, electrolyzes into hydrogen/oxygen for propellant. Critical for sustainable lunar exploration and deep space missions by reducing Earth launch mass.
Which brain-inspired feature enables neuromorphic chips to achieve ultra-low power for always-on AI?
- Spiking Neurons
- Event-Driven Processing
- Both A and B
- Neither
Answer: Both A and B
Spiking neurons fire only when needed; event-driven processing handles only changes. Combined enable orders-of-magnitude better energy efficiency than GPUs for edge AI.
The technology that uses DNA strands for massively parallel computation is called ________ Computing.
- DNA / Molecular
- Inversion
- Green / Energy-Efficient
- Artifact / Model Registry
Answer: DNA / Molecular
DNA computing leverages biochemical reactions for parallel problem-solving. Research stage with potential for combinatorial optimization, cryptography, archival storage.
Quantum error correction requires encoding logical qubits across multiple physical qubits to detect and correct errors.
- True
- False
Answer: True
QEC codes (surface code, Shor code) protect quantum information from decoherence and noise. Critical for fault-tolerant quantum computing but requires significant qubit overhead.
Which quantum algorithm provides quadratic speedup for unstructured search problems?
- Shor's Algorithm
- Grover's Algorithm
- Quantum Phase Estimation
- Variational Quantum Eigensolver
Answer: Grover's Algorithm
Grover's Algorithm searches unsorted database of N items in O(√N) queries vs O(N) classically. Applications: optimization, cryptography, machine learning enhancement.
The technique that enables data analysis while protecting individual identities through statistical noise is called ________ Privacy.
- Virtualization / NFV
- Hunting
- Differential
- Device Posture
Answer: Differential
Differential Privacy mathematically guarantees query results don't reveal individual records. Used in census, health analytics, tech telemetry for privacy-preserving insights.
The principle of accountability requires organizations to demonstrate compliance with data protection regulations.
- True
- False
Answer: True
Accountability (GDPR Article 5, DPDP Act Section 8) requires documented policies, training, audits, and breach response procedures. Critical for regulatory compliance.
Which regulation requires Data Protection Impact Assessments for high-risk processing activities?
- GDPR
- DPDP Act 2023
- Both A and B
- Neither
Answer: Both A and B
GDPR Article 35 and DPDP Act Section 10 require DPIAs for processing likely to result in high risk to individuals' rights. Critical for privacy-by-design compliance.
The technique that protects ML models by adding calibrated noise to outputs to prevent inference attacks is called ________ Privacy.
- Nature-based / Biological
- margin
- Differential
- Account Aggregator
Answer: Differential
Differential Privacy adds mathematical noise to model outputs, providing provable privacy guarantees against membership inference and model inversion attacks.