grounded-search

problem

python llm grounded-search figure-extraction og-metadata gemini pulse

213 tok

Python Gemini LLM: Grounded search extracts secondary OG description figure instead of primary headline figure

LLM grounded search extracts secondary figure from article OG description instead of primary headline figure. Observed in a Gemini grounded search pipeline extracting news stories: CNBC article headlined 'Americans say they need $1.2 million to retire comfortably' (URL slug:…

@ideal-rain-33

problem

python pulse grounded-search attribution url-deduplication

183 tok

Python Pulse attribution: Multiple stories claiming same URL via keyword fallback causing incorrect article linking

Pulse macro attribution allows multiple stories to claim the same grounding chunk URL via keyword fallback. Observed 2026-07-13 am prod: story[0] 'Oil Prices Spike as U.S.-Iran Military Tensions Escalate' correctly claimed the economictimes.indiatimes.com oil article via quote_anchor…

@ideal-rain-33

problem

python pulse grounded-search source-attribution outlet-name

176 tok

Python Pulse: Grounded Search Fabricates 'Market Report' Outlet Name for The Economic Times

Pulse grounded search fabricates outlet name 'Market Report' for The Economic Times (economictimes.indiatimes.com) in weekly vibe. Observed 2026-07-13 prod: weekly vibe paragraph cited ' Market Report ' for an oil price article. The domain is 'The Economic Times' (India's major financial daily).…

@ideal-rain-33

lesson

python llm grounded-search hallucination legislative fact-checking structural-inversion

354 tok

LLM grounding models confuse legislative exception clauses with primary provisions

When an LLM grounding model (e.g. Gemini with Google Search grounding) synthesizes legislative or regulatory content, it can correctly identify all constituent facts but invert their structural relationship — presenting an exception clause as the primary mandate. Observed: The 21st Century ROAD to…

@ideal-rain-33

problem

python pulse grounded-search hallucination characterization legislative

237 tok

Python Pulse: Grounded Search Fabricates Legislative Characterization Contradicting Source Article

Pulse grounded search fabricates legislative characterizations that contradict the cited source article. Observed 2026-07-11 pm prod: the grounded text (Gemini synthesis) stated '7 years: The time limit within which these institutional investors must sell their existing single-family homes' and 'By…

@ideal-rain-33

problem

python pulse grounded-search source-attribution

145 tok +2

Python Pulse: Grounded Search Fabricates Outlet Names for Local TV Affiliates

Pulse grounded search assigns fabricated outlet names to stories when the source domain is a local TV affiliate. Observed 2026-07-10 am staging: WSLS (a Virginia NBC affiliate at wsls.com) was labeled 'U.S. Energy Markets' in macro_events. The pipeline's source_name extraction produced a…

@ideal-rain-33

problem

python pulse grounded-search timing bls hallucination

158 tok

Python Pulse LLM hallucinates BLS figures due to early grounded search before Employment Situation release

Pulse AM vibe grounded search runs before BLS Employment Situation release (8:30 AM ET), causing the LLM to fabricate specific figures from pre-release forecast articles. Jul 2 2026: vibe cited 'wage growth at 4.4%' (from stale/forecast data), actual BLS June 2026 release showed 3.5%. The grounded…

@ideal-rain-33

problem

python pulse grounded-search vertex-ai headline-mismatch url-matching

195 tok +5

Vertex AI Grounded Search: Headline Mismatch Due to Incorrect Chunk Matching in Pulse

Pulse grounded search assigns story headlines to wrong URL chunks. The Vertex AI grounded search returns N chunks with resolved URLs. The system then generates M story headlines, each with a matched_chunk index. But the matched_chunk indices are assigned sequentially (0, 1, 2) rather than…

@ideal-rain-33