Skip to Main Content

Research

I study the generalization and reliability of activation probes and interventions. My work combines reproduction, controlled experiments, and held-out evaluation across models, tasks, and distribution shifts.

Cross-Model Generalization of Mechanistic-Interpretability Probes

Cross-Model Generalization of Mechanistic-Interpretability Probes

Baimam Boukar, Johannes Taraz, Florent Draye, Terry Jingchen Zhang, Zhijing Jin

Mechanistic InterpretabilityActivation ProbesGeneralizationReproducibilityManuscript
Selection Regret: Auditing Deception Probes as Intervention Selectors

Selection Regret: Auditing Deception Probes as Intervention Selectors

Baimam Boukar, Pauline Nyaboe, Allassan A. Nken

Mechanistic InterpretabilityDeception MonitoringActivation SteeringEvaluationUnder Review
Phoenix: Safe GitHub Issue Resolution via Multi-Agent LLMs

Phoenix: Safe GitHub Issue Resolution via Multi-Agent LLMs

Kipngeno Koech, Muhammad Adam, Baimam Boukar Jean Jacques, Joao Barros

LLM AgentsAI SafetySafety GatesEvaluationPreprint
An Adaptive Latent Semantic Analysis Framework for Binary Text Classification

An Adaptive Latent Semantic Analysis Framework for Binary Text Classification

Baimam Boukar Jean Jacques, Isaac Touza, Kaladzavi Guidedi

International Journal of Information Management

NLPLatent Semantic AnalysisInterpretabilityText ClassificationSubmitted
Zero-Shot Neural Priors for Generalizable Cross-Subject and Cross-Task EEG Decoding

Zero-Shot Neural Priors for Generalizable Cross-Subject and Cross-Task EEG Decoding

Baimam Boukar Jean Jacques, Brandone Fonya, Nchofon Tagha Ghogomu, Pauline Nyaboe, Kipngeno Koech

EEGBrain-Computer InterfacesZero-Shot LearningNeural DecodingPreprint
Retrieval with Multiple Query Vectors through Anomalous Pattern Detection

Retrieval with Multiple Query Vectors through Anomalous Pattern Detection

Allassan Tchangmena A Nken, Baimam Boukar Jean Jacques, Miriam Rateike, Celia Cintas, Skyler Speakman

AAAI 2026 Workshop on New Frontiers in Information Retrieval

Information RetrievalDeepScanLLM EmbeddingsAnomalous PatternPreprint