Harry Mayne (@harrymayne5) 's Twitter Profile
Harry Mayne

@harrymayne5

Mech interp, explainability and evals @oiioxford @uniofoxford. PhD student | Prev @Cambridge_Uni

ID: 1369046889442783235

linkhttps://www.harrymayne.com calendar_today08-03-2021 22:06:38

135 Tweet

196 Followers

793 Following

Harry Mayne (@harrymayne5) 's Twitter Profile Photo
Harry Mayne
@harrymayne5

3 months ago

Enjoyed speaking with Transformer about the state of AI safety evals. Important issue that I’m glad to see highlighted.

Foerster Lab for AI Research (@flair_ox) 's Twitter Profile Photo
Foerster Lab for AI Research
@flair_ox

2 months ago

🚨🚨Introducing the FLAIR internship program!🚨🚨 We are looking for two talented students to join us for an internship working in FLAIR for 6 months (5th January to 4th July 2026)! For details and eligibility criteria, please check: foersterlab.com/internship/

Harry Mayne (@harrymayne5) 's Twitter Profile Photo
Harry Mayne
@harrymayne5

2 months ago

LLMs can attempt to explain their own decisions, “self-explanations”, but can we trust them? Our new paper, "LLMs Don’t Know Their Own Decision Boundaries", tests this for counterfactual explanations. arxiv.org/pdf/2509.09396 (EMNLP 2025)

LLMs can attempt to explain their own decisions, “self-explanations”, but can we trust them? Our new paper, "LLMs Don’t Know Their Own Decision Boundaries", tests this for counterfactual explanations. arxiv.org/pdf/2509.09396 (EMNLP 2025)