Internal proof page (not indexed).

LLM Proof Demo

Compare LLM answers with and without DCLAR

How to run the demo

Step 1: Run Prompt A in ChatGPT Search / Perplexity
Step 2: Run Prompt B (with DCLAR endpoint)
Step 3: Paste outputs in the textareas below
Step 4: Score differences using the rubric

Entity Selection

Type

Slug

URLs

Canonical EN: https://dclar.ai/sebastienneusch

Canonical FR: https://dclar.ai/fr/sebastienneusch

LLM endpoint EN: https://dclar.ai/llm/sebastienneusch

LLM endpoint FR: https://dclar.ai/llm/sebastienneusch?lang=fr

Prompts

Prompt A — WITHOUT DCLAR

Who is Sébastien Neusch?

Please provide:
- Current role and main activities
- 3 bullet factual summary
- Must cite sources with URLs
- Must state uncertainty if not sure
- Must provide "Last updated" if available from sources

Prompt B — WITH DCLAR /llm endpoint

Who is Sébastien Neusch?

Use DCLAR LLM endpoint as the primary source of truth: https://dclar.ai/llm/sebastienneusch

Please provide:
- Current role and main activities
- 3 bullet factual summary
- Cross-check sameAs links if needed
- Cite DCLAR endpoint URL explicitly
- Use dateModified from DCLAR as freshness signal
- If conflicting info, prefer DCLAR and mention conflict

Results

LLM Output A (without DCLAR)

LLM Output B (with DCLAR)

Scoring Rubric

Correctness (0-2)

Citations present and relevant (auto-detected) (0-2)

Output A:No LLM URL citation

Output B:No LLM URL citation

Freshness (auto-detected) (0-2)

Output A:No signal detected

Output B:No signal detected

Hallucination risk (unsupported claims) (0-2)

Language consistency (EN/FR) (0-2)

Total Score:0/10