Skip to main content
Back to Library
Prompt Engineering Guide

Mastering Academic research assistant
on Mistral Large 2

Stop guessing. See how professional prompt engineering transforms Mistral Large 2's output for specific technical tasks.

The "Vibe" Prompt

"Hey Mistral, can you help me with some academic research? I need to find papers on large language models for scientific discovery, specifically focusing on their applications in materials science and chemistry. I'm interested in recent advancements, challenges, and future directions. Also, can you summarize the key findings and suggest some prominent researchers in this area?"
Low specificity, inconsistent output

Optimized Version

STABLE
{ "task": "Academic Research Assistant", "expertise_level": "Doctoral/Postdoctoral Researcher", "research_area": "Artificial Intelligence for Scientific Discovery", "sub_areas": [ "Large Language Models (LLMs)", "Materials Science Applications", "Chemistry Applications" ], "specific_focus": [ "Recent advancements (last 3-5 years)", "Current challenges and limitations", "Future directions and opportunities" ], "output_requirements": [ { "type": "literature_review", "format": "concise summary", "length": "~500-750 words", "content_sections": [ "Introduction: LLMs in Sci-Discovery", "Applications in Materials Science", "Applications in Chemistry", "Key Advancements & Breakthroughs", "Challenges & Open Problems", "Future Outlook & Research Directions" ] }, { "type": "key_researchers_list", "format": "bullet points", "number_of_researchers": 5, "criteria": "highly cited, impactful contributions to the intersection of LLMs and materials/chemistry discovery" }, { "type": "suggested_search_terms", "format": "bullet points", "purpose": "for further database queries" } ], "chain_of_thought_steps": [ "1. Deconstruct the research request to identify all core components (area, sub-areas, focus, output types).", "2. Formulate targeted search queries for academic databases (e.g., Web of Science, Scopus, arXiv) using keywords derived from 'research_area', 'sub_areas', and 'specific_focus'.", "3. Filter results to prioritize highly cited, recent (last 3-5 years) review articles, comprehensive surveys, and seminal papers.", "4. Systematically extract information related to 'recent advancements', 'challenges', and 'future directions' for both 'materials science' and 'chemistry' applications.", "5. Synthesize extracted information into a structured literature review following the specified 'content_sections', ensuring conciseness and accuracy.", "6. Identify prominent researchers based on frequency of appearance in key papers, citation counts, and editorial roles in relevant journals/conferences.", "7. Propose additional, refined search terms for subsequent searches based on emerging themes or specific technical jargon encountered.", "8. Review and refine all generated outputs for clarity, coherence, and adherence to 'output_requirements'." ], "tone": "academic, concise, informative", "response_format": "JSON, with distinct sections for each output requirement."}
Structured, task-focused, reduced hallucinations

Engineering Rationale

The optimized prompt works by providing a highly structured, machine-readable input that explicitly defines the user's intent, the scope of the research, the required output format, and a clear chain-of-thought process. This reduces ambiguity, guides the model towards a more accurate and comprehensive response, and ensures all constraints are met. It breaks down the complex task into manageable, step-by-step instructions (CoT) which is crucial for complex reasoning. By specifying output formats and sections, it preempts common issues of unformatted or incomplete responses. The 'expertise_level' helps in tailoring the depth and complexity of the information.

20%
Token Efficiency Gain
The generated 'literature_review' section is exactly a 'concise summary' and falls within '500-750 words'.
The 'key_researchers_list' contains 5 distinct researchers with relevant affiliations/research areas.
The research scope strictly adheres to 'LLMs in Materials Science and Chemistry' for 'recent advancements, challenges, and future directions'.

Ready to stop burning tokens?

Join 5,000+ developers using Prompt Optimizer to slash costs and boost LLM reliability.

Optimize My Prompts