Tootfinder

Today I spoke at Stortorget in Helsingborg on why Ukraine’s resilience and our support are crucial. Three years into the war, commitment remains strong i Sweden. I’m sharing my speech here - about resistance, freedom, and responsibility. Please share!
https://carlheath.se/on-r…

Huoltoasema pilasi pohjavettä Helsingissä #Helsinki

Huoltoasema pilasi pohjavettä Helsingissä
Teboilin omankin selvityksen mukaan Linnavuorentiellä Vartioharjussa sijaitseva huoltoasema on vuosia pilannut toiminnallaan maaperää ja pohjavettä.

Tammikuussa, ennen kuntavaaleja, Malmin lentoaseman ystävät ry kirjoitti "Malmin lentokentän tulevaisuus on nyt helsinkiläisten käsissä" (siis vaaleissa).
No kuinka kävi? Demarit ja vasemmistoliitto voittivat isosti, Kokoomus ja varsinkin Perussuomalaiset hävisivät.
Ei tainnut Malmi-asia olla kovin tärkeä helsinkiläisille.
Onkohan Malmin lentoaseman ystävät ry nyt sitten sitä mieltä että "Malmin lentokentän tulevaisuus oli helsinkiläisten käsissä ja he sanoiv…

Mitigating Geospatial Knowledge Hallucination in Large Language Models: Benchmarking and Dynamic Factuality Aligning
Shengyuan Wang, Jie Feng, Tianhui Liu, Dan Pei, Yong Li
https://arxiv.org/abs/2507.19586

Mitigating Geospatial Knowledge Hallucination in Large Language Models: Benchmarking and Dynamic Factuality Aligning
Large language models (LLMs) possess extensive world knowledge, including geospatial knowledge, which has been successfully applied to various geospatial tasks such as mobility prediction and social indicator prediction. However, LLMs often generate inaccurate geospatial knowledge, leading to geospatial hallucinations (incorrect or inconsistent representations of geospatial information) that compromise their reliability. While the phenomenon of general knowledge hallucination in LLMs has been w…

Idag talade jag på Stortorget i Helsingborg om varför Ukrainas motståndskraft och vårt stöd är avgörande. Tre år in i kriget är engagemanget starkt runt om i Sverige! Glädjande! Jag delar mitt tal här - om motstånd, frihet och ansvar. Dela gärna vidare!
https://carlheath.se/om-motstandskraft

HalluSegBench: Counterfactual Visual Reasoning for Segmentation Hallucination Evaluation
Xinzhuo Li, Adheesh Juvekar, Xingyou Liu, Muntasir Wahed, Kiet A. Nguyen, Ismini Lourentzou
https://arxiv.org/abs/2506.21546

HalluSegBench: Counterfactual Visual Reasoning for Segmentation Hallucination Evaluation
Recent progress in vision-language segmentation has significantly advanced grounded visual understanding. However, these models often exhibit hallucinations by producing segmentation masks for objects not grounded in the image content or by incorrectly labeling irrelevant regions. Existing evaluation protocols for segmentation hallucination primarily focus on label or textual hallucinations without manipulating the visual context, limiting their capacity to diagnose critical failures. In respon…

Agentic AI and Hallucinations
Engin Iyidogan, Ali I. Ozkes
https://arxiv.org/abs/2507.19183 https://arxiv.org/pdf/2507.19183…

Agentic AI and Hallucinations
We model a competitive market where AI agents buy answers from upstream generative models and resell them to users who differ in how much they value accuracy and in how much they fear hallucinations. Agents can privately exert effort for costly verification to lower hallucination risks. Since interactions halt in the event of a hallucination, the threat of losing future rents disciplines effort. A unique reputational equilibrium exists under nontrivial discounting. The equilibrium effort, and t…

🇺🇦 #NowPlaying on KEXP's #PacificNotions
The Green Kingdom:
🎵 Winter Dream
#TheGreenKingdom
https://heliosmusic.bandcamp.com/track/winter-dream

LSD — My Problem Child
#Psychiatry

FRED: Financial Retrieval-Enhanced Detection and Editing of Hallucinations in Language Models
Likun Tan, Kuan-Wei Huang, Kevin Wu
https://arxiv.org/abs/2507.20930 https://

FRED: Financial Retrieval-Enhanced Detection and Editing of Hallucinations in Language Models
Hallucinations in large language models pose a critical challenge for applications requiring factual reliability, particularly in high-stakes domains such as finance. This work presents an effective approach for detecting and editing factually incorrect content in model-generated responses based on the provided context. Given a user-defined domain-specific error taxonomy, we construct a synthetic dataset by inserting tagged errors into financial question-answering corpora and then fine-tune fou…