https://labs.iovstudio.kr/en/news/self-preference IOV LABS en 2026-06-06T21:00:00+09:00 The judge in the mirror: LLM evaluators favor their own family, but cannot see why https://labs.iovstudio.kr/en/news/llm-forecast IOV LABS en 2026-06-05T21:00:00+09:00 Fluency is not foresight: LLMs forecast the future worse than a coin https://labs.iovstudio.kr/en/news/capitulation-curve IOV LABS en 2026-06-03T20:00:00+09:00 A model that knows it is right does not fold, even when you swear it is wrong https://labs.iovstudio.kr/en/news/observer-effect IOV LABS en 2026-06-02T22:00:00+09:00 LLMs know when they're being tested, and GPT gets more honest because of it https://labs.iovstudio.kr/en/news/enterprise-ai-playbook IOV LABS en 2026-06-02T21:00:00+09:00 The tools are mature. ROI is decided by the control system, not the tool. https://labs.iovstudio.kr/en/news/ai-design-tells IOV LABS en 2026-06-02T19:00:00+09:00 The AI design look is not taste. It is a finite set of defaults, and we measured it. https://labs.iovstudio.kr/en/news/convergence-pressure IOV LABS en 2026-06-01T19:00:00+09:00 It is not AI assistance that homogenizes a culture. It is the loop. https://labs.iovstudio.kr/en/news/llm-judge-bench IOV LABS en 2026-05-30T19:00:00+09:00 We tested the AI that grades other AIs. It's reliable until the answer is a tie. https://labs.iovstudio.kr/en/news/election-forecast-model IOV LABS en 2026-05-30T10:00:00+09:00 We built a forecasting model for the 2026 local elections and we'll grade it https://labs.iovstudio.kr/en/news/research-archived-with-doi IOV LABS en 2026-05-29T19:00:00+09:00 IOV LABS enters the permanent scientific record: our work now has a DOI https://labs.iovstudio.kr/en/news/founder-orcid-id IOV LABS en 2026-05-29T18:00:00+09:00 IOV LABS founder joins the global research record with an ORCID iD https://labs.iovstudio.kr/en/news/korean-text-render-benchmark IOV LABS en 2026-05-29T15:00:00+09:00 We benchmarked how well image models draw Korean. One can't at all. https://labs.iovstudio.kr/en/news/gen-media-quality-gating IOV LABS en 2026-05-29T12:00:00+09:00 Can a cheap verifier gate generative media, and route it? https://labs.iovstudio.kr/en/news/verifiable-codegen-study IOV LABS en 2026-05-29T09:00:00+09:00 0x-lang: a token benchmark and a verifiable-codegen study https://labs.iovstudio.kr/en/news/iov-labs-founded IOV LABS en 2026-05-01 IOV LABS launches in Seoul as an open-source AI research lab https://labs.iovstudio.kr/ko/news/self-preference IOV LABS ko 2026-06-06T21:00:00+09:00 거울 속의 심사자: LLM 평가자는 제 가족을 편애하지만, 왜인지는 모른다 https://labs.iovstudio.kr/ko/news/llm-forecast IOV LABS ko 2026-06-05T21:00:00+09:00 유창함은 예측이 아니다: LLM은 미래를 동전보다 못 맞힌다 https://labs.iovstudio.kr/ko/news/capitulation-curve IOV LABS ko 2026-06-03T20:00:00+09:00 정답을 아는 모델은 당신이 아무리 우겨도 접지 않는다 https://labs.iovstudio.kr/ko/news/observer-effect IOV LABS ko 2026-06-02T22:00:00+09:00 LLM은 시험당하는 걸 안다, 그리고 GPT는 그 때문에 더 정직해진다 https://labs.iovstudio.kr/ko/news/enterprise-ai-playbook IOV LABS ko 2026-06-02T21:00:00+09:00 도구는 성숙했다. ROI를 가르는 건 도구가 아니라 통제 시스템이다. https://labs.iovstudio.kr/ko/news/ai-design-tells IOV LABS ko 2026-06-02T19:00:00+09:00 AI 같은 디자인은 취향이 아니다. 유한한 기본값의 집합이고, 우리는 그걸 측정했다. https://labs.iovstudio.kr/ko/news/convergence-pressure IOV LABS ko 2026-06-01T19:00:00+09:00 문화를 동질화하는 것은 AI 보조가 아니다. 루프다. https://labs.iovstudio.kr/ko/news/llm-judge-bench IOV LABS ko 2026-05-30T19:00:00+09:00 AI를 채점하는 AI를 시험했다. 정답이 있을 땐 믿을 만하다, 동점이 되기 전까지는. https://labs.iovstudio.kr/ko/news/election-forecast-model IOV LABS ko 2026-05-30T10:00:00+09:00 2026 지방선거 예측 모델을 만들었다, 그리고 결과로 채점받겠다 https://labs.iovstudio.kr/ko/news/research-archived-with-doi IOV LABS ko 2026-05-29T19:00:00+09:00 IOV LABS, 영구 과학 기록에 이름을 올리다: 우리의 작업에 DOI가 붙었다 https://labs.iovstudio.kr/ko/news/founder-orcid-id IOV LABS ko 2026-05-29T18:00:00+09:00 IOV LABS 설립자, ORCID iD로 글로벌 연구 기록에 합류 https://labs.iovstudio.kr/ko/news/korean-text-render-benchmark IOV LABS ko 2026-05-29T15:00:00+09:00 이미지 모델의 한글 렌더링을 측정했다. 한 모델은 아예 못 썼다. https://labs.iovstudio.kr/ko/news/gen-media-quality-gating IOV LABS ko 2026-05-29T12:00:00+09:00 싼 검증기로 생성형 미디어 품질을 거르고, 라우팅까지 할 수 있을까 https://labs.iovstudio.kr/ko/news/verifiable-codegen-study IOV LABS ko 2026-05-29T09:00:00+09:00 0x-lang, 토큰 벤치마크와 검증 가능한 코드생성 연구 공개 https://labs.iovstudio.kr/ko/news/iov-labs-founded IOV LABS ko 2026-05-01 AI 연구소 IOV LABS, 서울에서 오픈소스 연구소로 출범