Search Tips for Researchers
Virza’s search adapts to how you ask. These tips help you get the most relevant results for different research tasks.
Match your query to your intent
| If you want to… | Write your query like… | Why |
|---|---|---|
| Find a specific paper | Smith 2024 attention mechanisms | Short keyword queries trigger exact lexical matching, which is fast and precise |
| Explore a concept | How do transformer architectures improve clinical NLP? | Question-form queries activate semantic search + hypothetical document generation for deeper matching |
| Find supporting/contradicting evidence | Does intermittent fasting reduce cardiovascular risk? | Claim-form queries trigger evidence-seeking mode, surfacing supporting and contrasting results |
| Jump to a paper by DOI | 10.1038/s41586-024-07487-w | Paste the DOI directly. Virza matches it instantly |
Use the right search tab
| Tab | When to use |
|---|---|
| My Library | You know the paper is in your workspace. Looking for specific content, cross-referencing, finding related documents within your collection. |
| All | You want to see both your library and potentially relevant external papers. Good for identifying gaps. |
| Discover | You’re building your library. Looking for papers you don’t have yet. Literature review research. |
Scope search to a collection
When working within a specific research project:
- Navigate to the Search page
- Select the My Library tab
- Use the Collection filter to restrict results to a specific collection
This is especially useful for literature reviews where you want to search within a curated set of papers.
Combine search with filters
After your initial search, narrow results using filters:
- Publication year: focus on recent research (e.g., 2023–2026) or historical context
- Methodology: filter by research method (meta-analysis, RCT, cohort study, etc.)
- Has tables/figures: find data-rich papers with extractable evidence
Query length matters
- 1–3 words: treated as keyword search → exact lexical matching, no semantic expansion
- 4+ words: may trigger LLM-based intent classification → semantic matching, query expansion
- Questions (ending with ?): always trigger full semantic pipeline including hypothetical document generation
Tips for specific research tasks
Systematic literature review
- Start with a broad question on the Discover tab
- Save relevant papers to a “Review Candidates” collection
- Search within that collection using specific methodology or population filters
- Use the All tab periodically to check for papers you may have missed
Finding contradictory evidence
Use claim-form queries:
- “Does X increase Y?”
- “Evidence against Z”
- “Contradicting findings on W”
These activate evidence-seeking mode, which surfaces papers with opposing conclusions.
Cross-document analysis
- Upload all relevant papers to a collection
- Use workspace-scoped chat to ask cross-cutting questions
- The AI searches across all documents and synthesizes evidence, citing specific sources
Your results improve over time. Virza’s workspace-aware ranking learns from your research focus. Papers aligned with your workspace’s research trajectory are boosted in relevance.
Last updated on