Filter and classify AI research content for relevance. Use when processing raw content from Twitter, Substacks, blogs, or podcasts to determine if it's worth extracting claims from. Assigns relevance scores, topics, and author categories.
Sign in to like and favorite skills
Assess content for relevance to AI research intelligence gathering. Filter noise and classify what remains.
How relevant is this to understanding AI research progress, capabilities, limitations, or field direction?
| Score Range | Meaning | Examples |
|---|---|---|
| 0.0-0.3 | Not relevant | Personal updates, off-topic, promotional |
| 0.3-0.6 | Tangentially relevant | General tech news, adjacent topics |
| 0.6-0.8 | Relevant | Discusses AI research, capabilities, field |
| 0.8-1.0 | Highly relevant | Substantive claims, predictions, research insights |
Assign ONE primary topic:
scalingreasoningagentssafetyinterpretabilitymultimodalrlhfroboticsbenchmarksinfrastructurepolicygeneralotherWhat kind of content is this?
predictionresearch-hintopinionfactualcritiquemetanoiseDoes this contain actual claims, arguments, or insights?
Substantive examples:
Non-substantive examples:
Classify the author:
lab-researchercriticacademicindependentjournalistunknownReturn JSON:
{ "assessments": [ { "itemIndex": 0, "relevance": 0.85, "topic": "reasoning", "contentType": "research-hint", "isSubstantive": true, "authorCategory": "lab-researcher", "brief": "One sentence summary" } ] }