24 citations · 120 across the 27 of their papers we have counts for
28 papers
Gaming Without an Attacker: Benchmark Fingerprinting in LLM-Driven Search Under Selection Pressure
Víctor Gallego
Benchmarks for systems that are optimized against the evaluation signal measure something different from what they claim. We document this concretely in two GPU-kernel-optimization…
A Hybrid Nested Harness for Decoupling Structure and Parameters in LLM-Driven Optimization
Víctor Gallego
In evolutionary algorithms powered by language models, the LLM acts as a single operator that simultaneously updates structural components (like control flow) and continuous parame…
Metal-Sci: A Scientific Compute Benchmark for Evolutionary LLM Kernel Search on Apple Silicon
Víctor Gallego
We present Metal-Sci, a 10-task benchmark of scientific Apple Silicon Metal compute kernels spanning six optimization regimes (stencils, all-pairs in -body problems, multi-field…
Discovering Cooperative Pipelines: Autoresearch for Sequential Social Dilemmas
Víctor Gallego
We study two-level autoresearch for cooperation: an outer-loop AI agent autonomously redesigns the inner-loop pipeline of an LLM policy-synthesis system for multi-agent Sequential…
Discovering Agentic Safety Specifications from 1-Bit Danger Signals
Víctor Gallego
Can large language model agents discover hidden safety objectives through experience alone? We introduce EPO-Safe (Experiential Prompt Optimization for Safe Agents), a framework wh…
Beyond Scalar Rewards: Dense Feedback for LLM Policy Synthesis in Sequential Social Dilemmas
Víctor Gallego
We propose an LLM harness that generates code-based policy functions for multi-agent environments, evaluates them with self-play, and refines them using feedback from previous iter…
Distilling Feedback into Memory-as-a-Tool
Víctor Gallego
We propose a framework that amortizes the cost of inference-time reasoning by converting transient critiques into retrievable guidelines, through a file-based memory system and age…
Specification Self-Correction: Mitigating In-Context Reward Hacking Through Test-Time Refinement
Víctor Gallego
Language models (LMs) are susceptible to in-context reward hacking, where they exploit flaws in tainted or faulty written specifications or rubrics to achieve high scores without f…
Configurable Preference Tuning with Rubric-Guided Synthetic Data
Víctor Gallego
Models of human feedback for AI alignment, such as those underpinning Direct Preference Optimization (DPO), often bake in a singular, static set of preferences, limiting adaptabili…
MetaSC: Test-Time Safety Specification Optimization for Language Models
Víctor Gallego
We propose a novel dynamic safety framework that optimizes language model (LM) safety reasoning at inference time without modifying model weights. Building on recent advances in se…
Merging Improves Self-Critique Against Jailbreak Attacks
Victor Gallego
The robustness of large language models (LLMs) against adversarial manipulations, such as jailbreak attacks, remains a significant challenge. In this work, we propose an approach t…
Configurable Safety Tuning of Language Models with Synthetic Preference Data
Victor Gallego
State-of-the-art language model fine-tuning techniques, such as Direct Preference Optimization (DPO), restrict user control by hard-coding predefined behaviors into the model. To a…
Refined Direct Preference Optimization with Synthetic Data for Behavioral Alignment of LLMs
Víctor Gallego
In this paper, we introduce \emph{refined Direct Preference Optimization} (rDPO), a method for improving the behavioral alignment of Large Language Models (LLMs) without the need f…
Distilled Self-Critique of LLMs with Synthetic Data: a Bayesian Perspective
Victor Gallego
This paper proposes an interpretation of RLAIF as Bayesian inference by introducing distilled Self-Critique (dSC), which refines the outputs of a LLM through a Gibbs sampler that i…
Fast Adaptation with Bradley-Terry Preference Models in Text-To-Image Classification and Generation
Victor Gallego
Recently, large multimodal models, such as CLIP and Stable Diffusion have experimented tremendous successes in both foundations and applications. However, as these models increase…
ZYN: Zero-Shot Reward Models with Yes-No Questions for RLAIF
Victor Gallego
In this work, we address the problem of directing the text generation of a language model (LM) towards a desired behavior, aligning the generated text with the preferences of the h…
Personalizing Text-to-Image Generation via Aesthetic Gradients
Victor Gallego
This work proposes aesthetic gradients, a method to personalize a CLIP-conditioned diffusion model by guiding the generative process towards custom aesthetics defined by the user f…
Tuna-AI: tuna biomass estimation with Machine Learning models trained on oceanography and echosounder FAD data
Daniel Precioso, Manuel Navarro-García, Kathryn Gavira-O'Neill +4
Echo-sounder data registered by buoys attached to drifting FADs provide a very valuable source of information on populations of tuna and their behaviour. This value increases whent…
Contributions to Large Scale Bayesian Inference and Adversarial Machine Learning
Víctor Gallego
The rampant adoption of ML methodologies has revealed that models are usually adopted to make decisions without taking into account the uncertainties in their predictions. More cri…
Data sharing games
Víctor Gallego, Roi Naveiro, David Ríos Insua +1
Data sharing issues pervade online social and economic environments. To foster social progress, it is important to develop models of the interaction between data producers and cons…
Adversarial Risk Analysis (Overview)
David Banks, Víctor Gallego, Roi Naveiro +1
Adversarial risk analysis (ARA) is a relatively new area of research that informs decision-making when facing intelligent opponents and uncertain outcomes. It enables an analyst to…
Protecting Classifiers From Attacks
Victor Gallego, Roi Naveiro, Alberto Redondo +2
In multiple domains such as malware detection, automated driving systems, or fraud detection, classification algorithms are susceptible to being attacked by malicious agents willin…
Adversarial Machine Learning: Bayesian Perspectives
David Rios Insua, Roi Naveiro, Victor Gallego +1
Adversarial Machine Learning (AML) is emerging as a major field aimed at protecting machine learning (ML) systems against security threats: in certain scenarios there may be advers…
Variationally Inferred Sampling Through a Refined Bound for Probabilistic Programs
Victor Gallego, David Rios Insua
A framework to boost the efficiency of Bayesian inference in probabilistic programs is introduced by embedding a sampler inside a variational posterior approximation. We call it th…