Jonathan Cohen

Traian Rebedea, Razvan Dinu, Makesh Sreedhar et al.

NeMo Guardrails is an open-source toolkit for easily adding programmable guardrails to LLM-based conversational systems. Guardrails (or rails for short) are a specific way of controlling the output of an LLM, such as not talking about topics considered harmful, following a predefined dialogue path, using a particular language style, and more. There are several mechanisms that allow LLM providers and developers to add guardrails that are embedded into a specific model at training, e.g. using model alignment. Differently, using a runtime inspired from dialogue management, NeMo Guardrails allows developers to add programmable rails to LLM applications - these are user-defined, independent of the underlying LLM, and interpretable. Our initial results show that the proposed approach can be used with several LLM providers to develop controllable and safe LLM applications using programmable rails.

Zack Dulberg, Rachit Dubey, Isabel M. Berwian et al.

The problem of balancing conflicting needs is fundamental to intelligence. Standard reinforcement learning algorithms maximize a scalar reward, which requires combining different objective-specific rewards into a single number. Alternatively, different objectives could also be combined at the level of action value, such that specialist modules responsible for different objectives submit different action suggestions to a decision process, each based on rewards that are independent of one another. In this work, we explore the potential benefits of this alternative strategy. We investigate a biologically relevant multi-objective problem, the continual homeostasis of a set of variables, and compare a monolithic deep Q-network to a modular network with a dedicated Q-learner for each variable. We find that the modular agent: a) requires minimal exogenously determined exploration; b) has improved sample efficiency; and c) is more robust to out-of-domain perturbation.

Siddharth Suresh, Kushin Mukherjee, Tyler Giallanza et al.

Semantic feature norms have been foundational in the study of human conceptual knowledge, yet traditional methods face trade-offs between concept/feature coverage and verifiability of quality due to the labor-intensive nature of norming studies. Here, we introduce a novel approach that augments a dataset of human-generated feature norms with responses from large language models (LLMs) while verifying the quality of norms against reliable human judgments. We find that our AI-enhanced feature norm dataset, NOVA: Norms Optimized Via AI, shows much higher feature density and overlap among concepts while outperforming a comparable human-only norm dataset and word-embedding models in predicting people's semantic similarity judgments. Taken together, we demonstrate that human conceptual knowledge is richer than captured in previous norm datasets and show that, with proper validation, LLMs can serve as powerful tools for cognitive science research.

Marco Nurisso, Jesseba Fernando, Raj Deshpande et al.

Intelligent systems must deploy internal representations that are simultaneously structured -- to support broad generalization -- and selective -- to preserve input identity. We expose a fundamental limit on this tradeoff. For any model whose representational similarity between inputs decays with finite semantic resolution $\varepsilon$, we derive closed-form expressions that pin its probability of correct generalization $p_S$ and identification $p_I$ to a universal Pareto front independent of input space geometry. Extending the analysis to noisy, heterogeneous spaces and to $n>2$ inputs predicts a sharp $1/n$ collapse of multi-input processing capacity and a non-monotonic optimum for $p_S$. A minimal ReLU network trained end-to-end reproduces these laws: during learning a resolution boundary self-organizes and empirical $(p_S,p_I)$ trajectories closely follow theoretical curves for linearly decaying similarity. Finally, we demonstrate that the same limits persist in two markedly more complex settings -- a convolutional neural network and state-of-the-art vision-language models -- confirming that finite-resolution similarity is a fundamental emergent informational constraint, not merely a toy-model artifact. Together, these results provide an exact theory of the generalization-identification trade-off and clarify how semantic resolution shapes the representational capacity of deep networks and brains alike.