Skip to content
View hiroki-tamba-research's full-sized avatar

Block or report hiroki-tamba-research

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse

Hiroki Tamba | 丹波大樹

Independent researcher — AI evaluation infrastructure, narrative intelligence, LLM grader reliability.

Member of the Coalition for Independent Technology Research (CITR). Author of the Reproducible Grading section in the UK AISI inspect_ai documentation.

Recent contributions

ORCID: 0009-0004-7635-0741 · OSF

Licensing & Attribution

All original methodology, evaluation frameworks, empirical datasets, and preprints I publish across platforms (arXiv, Zenodo, figshare, OSF, GitHub, LinkedIn) are released under CC BY-NC-ND 4.0 unless otherwise noted, in accordance with academic norms of attribution, non-commercial exchange, and preservation of authorial voice.

  • Attribution required: cite the original work.
  • Non-commercial use: integration into for-profit products, services, or internal production evaluation pipelines requires prior written consent.
  • No derivatives: modification, adaptation, or re-implementation of the methodology in derivative frameworks requires prior written consent.

For commercial licensing, collaboration, or methodology consulting: hiroki.tamba@gmail.com

Popular repositories Loading

  1. model-routing-transparency model-routing-transparency Public

    Per-response model identity disclosure in frontier AI APIs: cross-vendor comparison (Google Gemini vs. Anthropic Claude, June 2026)

    Python 1

  2. refusal-bypass-by-routing refusal-bypass-by-routing Public

    Anchor deposit: content-class-triggered model substitution in Anthropic Claude Code (2026-07-10). 46 samples, N=15/15 deterministic on attack-mechanics prompts, transparent refusal under raw Messag…

    1 1

  3. japan-cannabis-act-its japan-cannabis-act-its Public

    Pre-registered interrupted time series analysis of drug enforcement patterns following Japan's December 2023 Cannabis Control Act amendment. OSF Registration: https://doi.org/10.17605/OSF.IO/S5JAQ

    R

  4. saluscope saluscope Public

    Multi-source health & development data explorer via open APIs. Extensible to any indicator.

    HTML

  5. strategic-narrative-terminal strategic-narrative-terminal Public

    Realtime geopolitical narrative monitoring interface.

  6. llm-judge-nondeterminism llm-judge-nondeterminism Public

    Empirical reproducibility note: non-determinism in LLM-as-judge graders (generalized from Japan-AISI/aisev #25)

    Python 1