Solving algorithmic tasks with many sequential steps typically requires scaling up transformer depth or chain-of-thought length, incurring unfavorable parameter and inference-compute costs, respectively. Looped transformers, which reuse one block of layers across many iterations, are poised to address this, since they...
Magnus Sesodia, Juil Sock, P. Torr et al.· 0 citations
Reinforcement learning with verifiable rewards (RLVR) has led to significant gains in math and code generation. Extending these gains to open-ended domains remains challenging: ground-truth verification is unavailable, human annotation is expensive, and learned reward models are prone to reward hacking. The generation-...
S. Motwani, Tejas Krishnan, Alistair Letcher et al.· 1 citation
This work investigates three such capabilities - persuasion, deception, and coercion - across five realistic evaluation environments and a variety of agents instantiated with frontier language models, observing significant inter-agent influence capabilities.
Chandler Smith, Qian-Xue Guo, Cecilia Tilli et al.· 0 citations
This work constructs a concrete prompt-injection attack that succeeds despite all constituent agents individually operating CaMeL, and introduces multi-CaMeL, an agent-to-agent communication protocol that preserves provenance across agent boundaries by separating trusted natural-language instructions from untrusted dat...
James Peters-Gill, Avi Semler, Henning Bartsch et al.· 1 citation
We study state-adversarial Markov decision processes (SA-MDPs) as games of observation-space attacks: at each step, an agent selects an action from a received observation while an adversary$\unicode{x2014}$who knows the true state the agent is in$\unicode{x2014}$chooses a perturbed observation within a state-dependent...
Multi-agent LLM systems are increasingly deployed for complex, long-horizon tasks or emerge as a natural consequence of agents interacting in the wild. Yet they give rise to significant safety and security risks: the flexible protocols that enable task generalization also expose novel threats, from cascading prompt inj...
Ben Hagag, William L. Anderson, S. Chakraborty et al.· arXiv.org· 1 citation
By repurposing Selective Gradient Masking, a suite of six 254M-parameter language models on English Wikipedia with graded levels of disentanglement between biology and non-biology knowledge are trained, it is found that more disentangled models consistently achieve better retain-forget trade-offs.
Evžen Wybitul, Tim G. J. Rudner, C. D. de Witt· 0 citations
It is argued that machine learning-based climate models should be designed to pass generalization tests to prevent overfitting on present-day regional climate.
Maren Höver, Milan Klöwer, Christian Schroeder de Witt et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
