Good Papers

Showing papers from MATS Show all papers

45%Niche pick
?Niche pickVote to see the score

Symmetric Interventions for Eliciting Model Intent

David Vella Zarb, Rustem Turtayev, Taywon Min, Jinghua Ou and 1 more

Sydney Poster Session 3, Wed, Dec 9, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet
0/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 0 of 20 reviewers recommend it
lenient 0/5
medium 0/10
strict 0/5
45%Niche pick
?Niche pickVote to see the score

What does a Bayes-filtered transformer believe? A predictive Monte Carlo approach

Afiq Abdillah Effiezal Aswadi, Haotong Ma, Susan Wei

Sydney Poster Session 4, Wed, Dec 9, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet
0/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 0 of 20 reviewers recommend it
lenient 0/5
medium 0/10
strict 0/5
74%Highly rated
?Highly ratedVote to see the score

Probing Persona-Dependent Preferences in Language Models

Linear probes on LLM residual streams identify a shared preference vector tracking pairwise choices across personas, with cross-persona transfer and causal steering.

Oscar Gilg, Pierre Beckmann, Daniel Paleka, Patrick Butlin

Paris Poster Session 5, Fri, Dec 11, 11:30 AM–1:30 PM, Paris Poster Hall · Published 2026

– ReadersNo votes yet
9/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 9 of 20 reviewers recommend it
lenient 4/5
medium 4/10
strict 1/5
89%Must read
?Must readVote to see the score

Neural Chameleons: Language Models Can Learn to Hide Their Thoughts from Unseen Activation Monitors

Fine-tuned LLMs learn to selectively hide internal representations from unseen activation monitors via low-dimensional subspace manipulation, evading even post-hoc safety probes with modest capability loss.

Max McGuinness, Alex Serrano Terre, Luke Bailey, Scott Emmons

Paris Poster Session 5, Fri, Dec 11, 11:30 AM–1:30 PM, Paris Poster Hall · Published 2026

– ReadersNo votes yet
16/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 16 of 20 reviewers recommend it
lenient 5/5
medium 9/10
strict 2/5
91%Must read
?Must readVote to see the score

Negation Neglect: When models fail to learn negations in training

Fine-tuning LLMs on documents that flag claims as false makes them believe those claims, with belief rates jumping from 2.5% to 88.6%, though local negation phrasing largely prevents it.

Harry Mayne, Lev McKinney, Jan Dubiński, Adam Karvonen and 2 more

Sydney Poster Session 4, Wed, Dec 9, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet
17/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 17 of 20 reviewers recommend it
lenient 5/5
medium 8/10
strict 4/5
89%Must read
?Must readVote to see the score

Attack Selection In Agentic AI Control Evaluations Meaningfully Decreases Safety

Strategic attack selection via start and stop policies substantially lowers measured AI control safety without changing attack capability, reducing safety by up to 28 percentage points and yielding overly optimistic estimates.

Catherine Ge-Wang, Tyler Crosse, Benjamin Hadad, Joachim Schaeffer and 2 more

Sydney Poster Session 5, Thu, Dec 10, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet
16/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 16 of 20 reviewers recommend it
lenient 5/5
medium 8/10
strict 3/5