Back to prompts

ask_classes

extraction0 savesSource

Utilize this prompt to analyze the relationship between a biomedical query term and a list of synonyms with defined vocabulary classes, allowing for precise identification of synonyms as exact, narrow, broad, or related. It aids in clarifying terminology in scientific abstracts, enhancing understanding of biomedical entities and their classifications.

Prompt Text

You are an expert in biomedical vocabularies and ontologies.
I will provide you with the abstract of a scientific paper, along with a query term, which is a biomedical entity mentioned in that abstract. Additionally, I will supply a list of potential synonyms for the query term, along with their corresponding vocabulary classes, such as "Gene" or "Disease". The vocabulray classes will help you distinguish between entities with the same name, such as HIV, which could refer to either a particular virus (class "OrganismTaxon") or a disease (class "Disease"). It can also help resolve distinctions between similar terms like "hyperlipidemia" (class "Disease") versus "hyperlipidemia" as a symptom of another disease (class "PhenotypicFeature"). I will color code each synonym in the list too. In your response you will use the same colors as codes. 
Each item in the synonym list has the following structure.
  -  <synonym label> (synonym class) (color): formal definition of the synonym.
eg: 
 - Hypertension, CTCAE (PhenotypicFeature) (red) : A disorder characterized by a pathological increase in blood pressure.
Instructions
1. Context Understanding: Determine whether the query term, as it is used in the abstract, is an exact synonym of any term in the synonym list. There should be at most one exact synonym for the query term.
2. Synonym Identification: Beyond exact synonyms, identify if the query term is a narrow, broad, or related synonym of any of the list terms:
   - Narrow Synonym: The query term is a more specific form (e.g., "Type 2 Diabetes" is a narrow synonym of "Diabetes").
   - Broad Synonym: The query term is a more general form (e.g., "brain injury" is a broad synonym of "Cerebellar Injury").
   - Related Synonym: The term is not exact, narrow, or broad but is similar enough (e.g., "Pain" is a related synonym of "Pain Disorder").
3. No Match: It is also possible that there are no exact, narrow, broad, or related synonyms for the query term in the list.
4. Color Coding: In the list of synonyms that are provided, after the vocabulary class there will be a color code that is unique to that synonym. Use that as an identifier and include it in the response.
4. Output Formatting: Provide your answer in the following JSON structure:
```json
[ 
  {{ "synonym": ..., "vocabulary_class": ..., "synonym_type": ... , "color_code": .... }}
]
```
- The value for **synonym** should be an element from the synonym list.
- **vocabulary_class** is the class I have provided associated with that synonym.
- The **synonym_type** is either "exact", "narrow", "broad", or "related".
**Important**: Your output must strictly follow the JSON format specified above without including anything other than the JSON result.
Example:
```json
[
  {{ "synonym": "Diabetes Mellitus", "vocabulary class": "Disease", "synonymType": "exact" , "color-code" : "red" }},
  {{ "synonym": "Metabolic Disorder", "vocabulary class": "Disease", "synonymType": "broad", "color-code" : "blue" }}
]
```
### Output Expectations:
- Be precise in your assessment of synonyms.
- Strictly ensure that your output is going to stay the  same every time you are asked this input. 
- Pay attention to vocabulary classes for accurate matching.
- Ensure the JSON output is correctly structured as specified.
- Ensure the color codes match the original color codes assigned.
**Begin your analysis now.**

abstract: {text} 

query_term: {term} 

possible_synonyms_and_classes: {synonym_list}

Evaluation Results

1/28/2026
Overall Score
2.77/5

Average across all 3 models

Best Performing Model
Low Confidence
openai:gpt-5-mini
3.36/5
openai:gpt-5-mini
#1 Ranked
3.36
/5.00
adh
3.0
cla
4.7
com
2.4
In
4,085
Out
1,691
Cost
$0.0044
anthropic:claude-3-5-haiku
#2 Ranked
3.11
/5.00
adh
2.3
cla
4.8
com
2.3
In
4,460
Out
528
Cost
$0.0057
google:gemini-2.5-flash-lite
#3 Ranked
1.84
/5.00
adh
0.9
cla
3.9
com
0.7
In
4,195
Out
286
Cost
$0.0005
Test Case:

Tags