Aaron Mueller
@amuuueller.bsky.social
Postdoc at Northeastern and incoming Asst. Prof. at Boston U. Working on NLP, interpretability, causality. Previously: JHU, Meta, AWS
Reposted by Aaron Mueller
I also want to mention that the lang x computation research community at BU is growing in an exciting direction, especially with new faculty like @amuuueller.bsky.social, @anthonyyacovone.bsky.social, @nsaphra.bsky.social, &
@profsophie.bsky.social! Also, Boston is quite nice :)
@profsophie.bsky.social! Also, Boston is quite nice :)
November 19, 2025 at 5:20 PM
I also want to mention that the lang x computation research community at BU is growing in an exciting direction, especially with new faculty like @amuuueller.bsky.social, @anthonyyacovone.bsky.social, @nsaphra.bsky.social, &
@profsophie.bsky.social! Also, Boston is quite nice :)
@profsophie.bsky.social! Also, Boston is quite nice :)
Check out the paper and our demo features!
📜 Preprint: arxiv.org/abs/2511.01836
🧠 Play with temporal feature analysis on Neuronpedia: www.neuronpedia.org/gemma-2-2b/1...
📜 Preprint: arxiv.org/abs/2511.01836
🧠 Play with temporal feature analysis on Neuronpedia: www.neuronpedia.org/gemma-2-2b/1...
Priors in Time: Missing Inductive Biases for Language Model Interpretability
Recovering meaningful concepts from language model activations is a central aim of interpretability. While existing feature extraction methods aim to identify concepts that are independent directions,...
arxiv.org
November 14, 2025 at 3:48 PM
Check out the paper and our demo features!
📜 Preprint: arxiv.org/abs/2511.01836
🧠 Play with temporal feature analysis on Neuronpedia: www.neuronpedia.org/gemma-2-2b/1...
📜 Preprint: arxiv.org/abs/2511.01836
🧠 Play with temporal feature analysis on Neuronpedia: www.neuronpedia.org/gemma-2-2b/1...
I'm glossing over our deeper motivations from neuroscience (predictive coding) and linguistics here, but we believe there's significant cross-field appeal for those interested in intersections of cog sci, neuroscience, and machine learning!
November 14, 2025 at 3:48 PM
I'm glossing over our deeper motivations from neuroscience (predictive coding) and linguistics here, but we believe there's significant cross-field appeal for those interested in intersections of cog sci, neuroscience, and machine learning!
Jeff Elman famously showed us in 1990 that time is a rich signal in itself.
Our work demonstrates that this lesson applies equally well to interpretability methods. The inductive biases of interp methods should reflect the structure of what is being studied.
Our work demonstrates that this lesson applies equally well to interpretability methods. The inductive biases of interp methods should reflect the structure of what is being studied.
November 14, 2025 at 3:48 PM
Jeff Elman famously showed us in 1990 that time is a rich signal in itself.
Our work demonstrates that this lesson applies equally well to interpretability methods. The inductive biases of interp methods should reflect the structure of what is being studied.
Our work demonstrates that this lesson applies equally well to interpretability methods. The inductive biases of interp methods should reflect the structure of what is being studied.
TFA is designed to capture context-sensitive information. Consider parsing: SAE often assign each word to its most frequent syntactic category, regardless of context.
Meanwhile, TFA recovers the correct parse given the context!
Meanwhile, TFA recovers the correct parse given the context!
November 14, 2025 at 3:48 PM
TFA is designed to capture context-sensitive information. Consider parsing: SAE often assign each word to its most frequent syntactic category, regardless of context.
Meanwhile, TFA recovers the correct parse given the context!
Meanwhile, TFA recovers the correct parse given the context!
Aruna Sankaranarayanan, @arnabsensharma.bsky.social absensharma.bsky.social @ericwtodd.bsky.social ky.social @davidbau.bsky.social u.bsky.social @boknilev.bsky.social (2/2)
October 1, 2025 at 2:03 PM
Aruna Sankaranarayanan, @arnabsensharma.bsky.social absensharma.bsky.social @ericwtodd.bsky.social ky.social @davidbau.bsky.social u.bsky.social @boknilev.bsky.social (2/2)
Thanks again to the co-authors! Such a wide survey required a lot of perspectives. @jannikbrinkmann.bsky.social Millicent Li, Samuel Marks, @koyena.bsky.social @nikhil07prakash.bsky.social @canrager.bsky.social (1/2)
October 1, 2025 at 2:03 PM
Thanks again to the co-authors! Such a wide survey required a lot of perspectives. @jannikbrinkmann.bsky.social Millicent Li, Samuel Marks, @koyena.bsky.social @nikhil07prakash.bsky.social @canrager.bsky.social (1/2)
See the paper for more details! arxiv.org/abs/2408.01416
The Quest for the Right Mediator: Surveying Mechanistic Interpretability Through the Lens of Causal Mediation Analysis
Interpretability provides a toolset for understanding how and why neural networks behave in certain ways. However, there is little unity in the field: most studies employ ad-hoc evaluations and do not...
arxiv.org
October 1, 2025 at 2:03 PM
See the paper for more details! arxiv.org/abs/2408.01416
We also made the causal graph formalism more precise. Interpretability and causality are intimately linked; the latter makes the former more trustworthy and rigorous. This formal link should be strengthened in future work.
October 1, 2025 at 2:03 PM
We also made the causal graph formalism more precise. Interpretability and causality are intimately linked; the latter makes the former more trustworthy and rigorous. This formal link should be strengthened in future work.
One of the bigger changes was establishing criteria for success in interpretability. What units of analysis should you use if you know what you’re looking for? If you *don’t* know what you’re looking for?
October 1, 2025 at 2:03 PM
One of the bigger changes was establishing criteria for success in interpretability. What units of analysis should you use if you know what you’re looking for? If you *don’t* know what you’re looking for?
We still have a lot to learn in editing NN representations.
To edit or steer, we cannot simply choose semantically relevant representations; we must choose the ones that will have the intended impact. As @peterbhase.bsky.social found, these are often distinct.
To edit or steer, we cannot simply choose semantically relevant representations; we must choose the ones that will have the intended impact. As @peterbhase.bsky.social found, these are often distinct.
May 27, 2025 at 5:07 PM
We still have a lot to learn in editing NN representations.
To edit or steer, we cannot simply choose semantically relevant representations; we must choose the ones that will have the intended impact. As @peterbhase.bsky.social found, these are often distinct.
To edit or steer, we cannot simply choose semantically relevant representations; we must choose the ones that will have the intended impact. As @peterbhase.bsky.social found, these are often distinct.
By limiting steering to output features, we recover >90% of the performance of the best supervised representation-based steering methods—and at some locations, we outperform them!
May 27, 2025 at 5:07 PM
By limiting steering to output features, we recover >90% of the performance of the best supervised representation-based steering methods—and at some locations, we outperform them!
We define the notion of an “output feature”, whose role is to increase p(some token(s)). Steering these gives better results than steering “input features”, whose role is to attend to concepts in the input. We propose fast methods to sort features into these categories.
May 27, 2025 at 5:07 PM
We define the notion of an “output feature”, whose role is to increase p(some token(s)). Steering these gives better results than steering “input features”, whose role is to attend to concepts in the input. We propose fast methods to sort features into these categories.
Reposted by Aaron Mueller
Couldn’t be happier to have co-authored this will a stellar team, including: Michael Hu, @amuuueller.bsky.social, @alexwarstadt.bsky.social, @lchoshen.bsky.social, Chengxu Zhuang, @adinawilliams.bsky.social, Ryan Cotterell, @tallinzen.bsky.social
May 12, 2025 at 3:48 PM
Couldn’t be happier to have co-authored this will a stellar team, including: Michael Hu, @amuuueller.bsky.social, @alexwarstadt.bsky.social, @lchoshen.bsky.social, Chengxu Zhuang, @adinawilliams.bsky.social, Ryan Cotterell, @tallinzen.bsky.social
... Jing Huang, Rohan Gupta, Yaniv Nikankin, @hadasorgad.bsky.social, Nikhil Prakash, @anja.re, Aruna Sankaranarayanan, Shun Shao, @alestolfo.bsky.social, @mtutek.bsky.social, @amirzur, @davidbau.bsky.social, and @boknilev.bsky.social!
April 23, 2025 at 6:15 PM
... Jing Huang, Rohan Gupta, Yaniv Nikankin, @hadasorgad.bsky.social, Nikhil Prakash, @anja.re, Aruna Sankaranarayanan, Shun Shao, @alestolfo.bsky.social, @mtutek.bsky.social, @amirzur, @davidbau.bsky.social, and @boknilev.bsky.social!
This was a huge collaboration with many great folks! If you get a chance, be sure to talk to Atticus Geiger, @sarah-nlp.bsky.social, @danaarad.bsky.social, Iván Arcuschin, @adambelfki.bsky.social, @yiksiu.bsky.social, Jaden Fiotto-Kaufmann, @talhaklay.bsky.social, @michaelwhanna.bsky.social, ...
April 23, 2025 at 6:15 PM
This was a huge collaboration with many great folks! If you get a chance, be sure to talk to Atticus Geiger, @sarah-nlp.bsky.social, @danaarad.bsky.social, Iván Arcuschin, @adambelfki.bsky.social, @yiksiu.bsky.social, Jaden Fiotto-Kaufmann, @talhaklay.bsky.social, @michaelwhanna.bsky.social, ...
We’re eager to establish MIB as a meaningful and lasting standard for comparing the quality of MI methods. If you’ll be at #ICLR2025 or #NAACL2025, please reach out to chat!
📜 arxiv.org/abs/2504.13151
📜 arxiv.org/abs/2504.13151
MIB: A Mechanistic Interpretability Benchmark
How can we know whether new mechanistic interpretability methods achieve real improvements? In pursuit of meaningful and lasting evaluation standards, we propose MIB, a benchmark with two tracks spann...
arxiv.org
April 23, 2025 at 6:15 PM
We’re eager to establish MIB as a meaningful and lasting standard for comparing the quality of MI methods. If you’ll be at #ICLR2025 or #NAACL2025, please reach out to chat!
📜 arxiv.org/abs/2504.13151
📜 arxiv.org/abs/2504.13151