
Natural Language Understanding Team Deputy Director @RIKEN_AIP, Associate Professor @TohokuNLP, based in Sendai, Japan. PhD @UniHeidelberg @HITStudies
Sendai, Japan.
Joined January 2017
- Tweets172
- Following210
- Followers410
- Likes684
Pinned Tweet
We've released a new version of our pretained byte pair embeddings in 275 languages. Now with pip install, automatic download of embeddings and sentencepiece models, convenient subword segmentation, and tons of pretty UMAP visualizations. nlp.h-its.org/bpemb
Benjamin Heinzerling retweeted
Thanks a lot UncertaiNLP for the spotlight 😍
The workshop was really cool, and a great ending to EMNLP
Birds of a feather on Doing NLP research in Japan @emnlpmeeting starting in 5 minutes. Join if you want to chat with people working as researchers in Japan or interested in doing so. Room is G12
2
230
Benjamin Heinzerling retweeted
Excited to share my first postdoc paper with @SheffieldNLP ! 🤩
In this work we argue that supervised uncertainty quantification (UQ) needs better evaluation
Want to know more? Here's a little summary 🧵
Benjamin Heinzerling retweeted
This work has been really fun to work on! Massive thank you my fantastic collaborators @OrgadHadas @inuikentaro @benbenhh and @NafiseSadat
You can find the paper here:
arxiv.org/abs/2604.11662
Benjamin Heinzerling retweeted
#NLP2026
明日 3/11(水)11:15–12:45 @ B会場(B6-16)でポスター発表します!
Japan ⊂ Eastern Asia ⊂ Asia ... のような概念の階層構造が,⾔語モデルの中間層にどのように・どの程度埋め込まれているかについて明らかにしました。
anlp.jp/proceedings/annual_m…
Benjamin Heinzerling retweeted
Benjamin Heinzerling @benbenhh さんの『人工知能』のコラム、10名の博士課程学生が一斉に雇用されるGraduiertenkolleg(Research Training Group)というドイツの仕組みを紹介していて、とても面白い。jstage.jst.go.jp/article/jjs…
Benjamin Heinzerling retweeted
The deadline for AACL-IJCNLP tutorial proposals is on August 14th AoE. We are interested in introductory as well as cutting edge half/full day tutorials. Introductory tutorials can be non-CL/non-NLP but interesting to the community.
2025.aaclnet.org/calls/tutor…
Benjamin Heinzerling retweeted
Super excited to have a cool post-doc lined up for after my PhD, working on Uncertainty Quantification with @NafiseSadat at @SheffieldNLP, and @benbenhh and @inuikentaro from RIKEN, Japan.
Could not be happier 🙂
Benjamin Heinzerling retweeted
「言語モデルの内部機序:解析と解釈」のスライドを公開しました
speakerdeck.com/eumesy/analy…
言語処理学会年次大会 #NLP2025 のチュートリアル資料です。w/ @benbenhh さん @goro_koba さん
皆で言語モデルの中を凝視していきましょう🔬
3月10日(月) に言語モデルの解釈性のチュートリアルをします🔥
@tohoku_nlp の @benbenhh さん @goro_koba さんと一緒に、モデルの内部機序(内部表象・計算メカニズム)について既存のどんな資料よりも詳しく分かりやすく解説します @ #NLP2025 長崎/オンライン。お楽しみに〜!
Benjamin Heinzerling retweeted
3月10日(月) に言語モデルの解釈性のチュートリアルをします🔥
@tohoku_nlp の @benbenhh さん @goro_koba さんと一緒に、モデルの内部機序(内部表象・計算メカニズム)について既存のどんな資料よりも詳しく分かりやすく解説します @ #NLP2025 長崎/オンライン。お楽しみに〜!
#NLP2025 では、3/10(月)午後 に、4件のチュートリアルを行います!
言語モデルの内部機序:解析と解釈(仮)
Benjamin Heinzerling 先生(理化学研究所/東北大学)
横井 祥 先生(東北大学/理化学研究所)
小林 悟郎 先生(東北大学/理化学研究所)
anlp.jp/nlp2025/#tutorial
Benjamin Heinzerling retweeted
🚀 Reminder! 🚀
The deadline for the Postdoc in Uncertainty Quantification for Foundation Models at @sheffieldNLP with @NafiseSadat @benbenhh @inuikentaro is March 2nd—just a few days left! 🗓️
📆 Initial contract: 12 months (with a possibility of an additional 2-year extension)
Benjamin Heinzerling retweeted
🚀 Postdoc Opportunity! 🚀
With @benbenhh & in collaboration with @inuikentaro, we’re hiring a Postdoc at @sheffieldNLP for Uncertainty Quantification in Foundation Models, with a chance to spend time at RIKEN, Japan!
📅 Deadline: 2 March
🔗 More info: jobs.ac.uk/job/DLS026/resear…
Benjamin Heinzerling retweeted
#NLP2025 では、3/10(月)午後 に、4件のチュートリアルを行います!
言語モデルの内部機序:解析と解釈(仮)
Benjamin Heinzerling 先生(理化学研究所/東北大学)
横井 祥 先生(東北大学/理化学研究所)
小林 悟郎 先生(東北大学/理化学研究所)
anlp.jp/nlp2025/#tutorial
Benjamin Heinzerling retweeted
📢 Academic Job Alert: Deadline Approaching Fast – Only 2 Days Left! (Apply by November 30th)
I'm currently seeking 2 postdoctoral researchers and 3 PhD students to join my team at IT:U-Interdisciplinary Transformation University Austria (@itu_austria)!
LLMs give the right answer to questions like "When was Karl Popper born?" but do they have the right internal representation? We find evidence that they do: There are directions in activation space that have a causal, interpretable effect on the expression of numeric properties
By activation patching along these directions we can make the LM express earlier/smaller and larger/later values across all properties and LMs we looked at. This suggests that monotonic representation of numeric properties consistently emerges during pretraining



















