Danny To Eun Kim
@teknology.bsky.social
Solution2️⃣: Human-Elicitation
We designed an interface with visual prompts to induce a TOT state in human participants. Their queries closely match authentic TOT queries and captures genuine TOT experiences in a controlled setting.
We designed an interface with visual prompts to induce a TOT state in human participants. Their queries closely match authentic TOT queries and captures genuine TOT experiences in a controlled setting.
March 5, 2025 at 1:35 AM
Solution2️⃣: Human-Elicitation
We designed an interface with visual prompts to induce a TOT state in human participants. Their queries closely match authentic TOT queries and captures genuine TOT experiences in a controlled setting.
We designed an interface with visual prompts to induce a TOT state in human participants. Their queries closely match authentic TOT queries and captures genuine TOT experiences in a controlled setting.
Solution1️⃣: LLM-Elicitation
We built a TOT user simulator to produce synthetic queries. Results show high system rank correlation and linguistic similarity compared to real queries. This scalable simulated evaluation method overcomes data scarcity by simulating new queries on demand.
We built a TOT user simulator to produce synthetic queries. Results show high system rank correlation and linguistic similarity compared to real queries. This scalable simulated evaluation method overcomes data scarcity by simulating new queries on demand.
March 5, 2025 at 1:35 AM
Solution1️⃣: LLM-Elicitation
We built a TOT user simulator to produce synthetic queries. Results show high system rank correlation and linguistic similarity compared to real queries. This scalable simulated evaluation method overcomes data scarcity by simulating new queries on demand.
We built a TOT user simulator to produce synthetic queries. Results show high system rank correlation and linguistic similarity compared to real queries. This scalable simulated evaluation method overcomes data scarcity by simulating new queries on demand.