Max Kirkby

Primary

  1. Can a Language Model Continually Learn Facts in its Weights? — O'Neill, Kirkby, Liu and Psenka. CL4FMAgents Workshop, NeurIPS (2026)
  2. Fitting Vision Adapters at Frontier Scales — Lee, Partridge, Jayasekara, O'Neill, Kirkby and Psenka. VLM4RWD Workshop, NeurIPS (2026)
  3. The Optimization Landscape of Learning Compacted Context Models — Villeneuve, Sandomirsky, O'Neill, Kirkby and Psenka. CL4FMAgents Workshop, NeurIPS (2026)
  4. Still: Amortized KV Cache Compaction in a Single Forward Pass — O'Neill, Sandomirsky, Partridge, Jayasekara and Kirkby. arxiv (2026)
  5. A Single Direction of Truth: An Observer Model's Linear Residual Probe Exposes and Steers Contextual Hallucinations — O'Neill, Chalnev, Kirkby, Zhao and Jayasekara. Actionable Interpretability Workshop, ICML (2025)
  6. Resurrecting the Salmon: Rethinking Mechanistic Interpretability with Domain-Specific Sparse Autoencoders — O'Neill, Jayasekara and Kirkby. arxiv (2025)
  7. Guanylate-binding proteins: mechanisms of pattern recognition and antimicrobial functions — Kirkby et al. Trends in Biochemical Sciences (2023)
  8. Streptococcus makes the cut: Gasdermin A-induced pyroptosis — Zhao, Kirkby and Man. Cell Host & Microbe Immunology (2022)
  9. Running to save sight: The effects of exercise on retinal health and function — Chu-Tan, Kirkby and Natoli. Clinical & Experimental Ophthalmology (2021)

Co-Authored

  1. Biological, Radiological, and Clinical Significance of Spotty Calcification in Coronary Artery Disease — Fathieh, Gray, Kirkby et al. JACC: Cardiovascular Imaging (2026)
  2. Exploring Differences in Functional Connectivity in Australian Rules Football Players : A Resting-State fMRI Study on the Default Mode Network — Tran et al. biorxiv (2024)
  3. Four-Dimensional Flow MRI for Cardiovascular Evaluation (4DCarE): A Prospective Non-Inferiority Study of a Rapid Cardiac MRI Exam: Study Protocol and Pilot Analysis — Qin et al. Diagnostics (2024)
  4. Inflammasome protein scaffolds the DNA damage complex during tumor development — Shen et al. Nature Immunology (2024)
  5. Immunity against Moraxella catarrhalis requires guanylate‐binding proteins and caspase‐11‐NLRP3 inflammasomes — Enosi-Tuipulotu et al. EMBO Journal (2023)
  6. Voluntary exercise modulates pathways associated with amelioration of retinal degenerative diseases — Chu-Tan et al. Frontiers in Physiology (2023)
  7. Molecular mechanisms activating the NAIP‐NLRC4 inflammasome: implications in infectious disease, autoinflammation, and cancer — Kay et al. Immunological Reviews (2020)

Conference Proceedings

  1. Representation of a hierarchical goal-directed behaviour in medial frontal cortex — Kirkby et al. Computational and Systems Neuroscience (COSYNE) (2026)
  2. A spatial navigation paradigm for investigating hierarchical planning — Kirkby et al. Reinforcement Learning and Decision Making (RLDM) (2025)
  3. Multi-modal detection of changes in MND for the quantitative evaluation of response to 3K3A-APC treatment — Kirkby et al. 34th International Symposium on ALS/MND (2023)

Artefacts

  1. When does distillation help reinforcement learning? — O'Neill, Psenka and Kirkby. Base Labs Checkpoint (2026)
  2. Distillation without the dark — Kirkby and O'Neill. Baseten Research (2026)
  3. Dense, on-policy, or both? — Kirkby and O'Neill. Baseten Research (2026)
  4. Practical LoRA Research — Kirkby and O'Neill. Parsed Research (2025)
  5. Training loss predicts evaluation performance, even for non-verifiable tasks — O'Neill and Kirkby. Parsed Research (2025)
  6. Iterative SFT (iSFT): dense reward learning — O'Neill, Liu, Partridge and Kirkby. Parsed Research (2025)
  7. Lumina: building self-improving evaluation through customer-in-the-loop refinement — O'Neill, Partridge, Kirkby, Liu and Stefanopoulos. Parsed Research (2025)
  8. Write small, learn forever: rank-1 LoRA for continual learning — O'Neill, Liu, Kirkby and Partridge. Parsed Research (2025)
  9. Attention-based attribution: what your model is actually looking at — O'Neill, Liu, Canavan, Kirkby and Jayasekara. Parsed Research (2025)
  10. Fine-tuning small open-source LLMs to outperform large closed-source models by 60% on specialized tasks — O'Neill, Jayasekara and Kirkby. Parsed Research (2025)