# Liquid _Labs._

Research is a founding principle of Liquid AI. Our origins stem from over a decade of MIT research; where Liquid Neural Networks created efficient, multimodal architectures for our Liquid Foundation Models.

Liquid Labs carries this legacy forward as our dedicated team driving fundamental breakthroughs in the science of building intelligent, personalized, and adaptive machines.

## LongevityBench: An open benchmark and language models for AI in aging biology  
By Alex Zhavoronkov, Vladimir Naumov, Denis Sidorenko, Alex Aliper, Vladimir Aladinskiy, Ramin Hasani, Alexander Amini, Katerina Nasto, Mathieu Reymond, Rim Shayakhmetov, Zulfat Miftakhutdinov, and 2 others  
We introduce LongevityBench, an open suite of 17 tasks spanning five biodata domains, and use it to assess 18 frontier AI systems from six developer teams.

## In-Place Tokenizer Expansion for Pre-trained LLMs  
By Jimmy T.H. Smith, Tarek Dakhran, Alberto Cabrera, Simon S. Lee, Paul Pak, Aditya Tadimeti, Tim Seyde, Maxime Labonne, Alexander Amini, and Mathias Lechner  
New recipe expands pre-trained LLM tokenizers in place, cutting non-English token counts up to 4x with no quality loss.

## Fixed-Point Reasoners: Stable and Adaptive Deep Looped Transformers  
By Sajad Movahedi, Vera Milovanović, Shlomo Libo Feigin, Alexander Theus, Thomas Hofmann, Valentina Boeva, T. Konstantin Rusch, and Antonio Orvieto  
FPRM is a Transformer-based Fixed-Point Reasoning Model for looped architectures, using pre-norm layers, residual scaling, and fixed-point halting to improve signal propagation, adapt compute to task difficulty, and deliver strong results on Sudoku, Maze, state-tracking, and ARC-AGI.

## Neural Low-Discrepancy Sequences  
By Michael Etienne Van Huffel, Nathan Kirk, Makram Chahine, Daniela Rus, and T. Konstantin Rusch  
NeuroLDS introduces the first machine learning framework for finite low-discrepancy sequences, generating point sequences with low prefix discrepancy and outperforming classical LDS methods across numerical integration, robot motion planning, scientific machine learning, and simulation.

## SwitchCraft: A Programmatic Framework for Designing State-Switching Proteins  
By Bowen Jing, Mihir Bafna, Anisha Parsan, Heyuan Michael Ni, David Kwabi-Addo, Bryan Bryson, Adam Klivans, and Bonnie Berger  
SwitchCraft is a programmable framework for designing state-switching proteins using backpropagation through compositional constraints and structure prediction models, enabling allosteric regulation, ligand discrimination, and de novo fluorescent biosensor design for biotechnology.

## MeMo: Model as a Memory  
By Ryan Wei Heng Quek, Sanghyuk Lee, Alfred Wei Lun Leong, Arun Verma, Alok Prakash, Nancy F. Chen, Bryan Kian Hsiang Low, Daniela Rus, and Armando Solar-Lezama  
MeMo (Memory as a Model) is a modular framework that adds timely, domain-specific knowledge to LLMs without changing model weights, capturing cross-document relationships, resisting retrieval noise, avoiding catastrophic forgetting, and supporting plug-and-play use with open and closed-source models.

## Preconditioned DeltaNet: Curvature-aware Sequence Modeling for Linear Recurrences  
By Neehal Tumma, Noel Loo, and Daniela Rus  
Preconditioned delta-rule recurrences improve subquadratic long-context modeling by incorporating curvature from online least squares. The method links linear attention and the delta rule, introduces efficient variants of DeltaNet, GDN, and KDA, and boosts recall and language modeling performance.

## The Key to State Reduction in Linear Attention: A Rank-based Perspective  
By Philipp Nazari and T. Konstantin Rusch  
Linear attention offers a computationally efficient yet expressive alternative to softmax attention, maintaining a recurrent state that functions as a linear associative memory. However, recent empirical results indicate that the associative memory of trained linear attention models often exhibits a low-rank structure, suggesting that these models underexploit their capacity in practice.

## Low-Pass Flow Matching  
By Francesco M. Ruscio and T. Konstantin Rusch  
Flow Matching typically relies on white noise sources, a choice often misaligned with the power spectra of natural data, which tend to decay with frequency. To address this, we introduce a variant of Flow Matching based on an operator-modulated interpolant.

## AlphaQ: Calibration-Free Bit Allocation for Mixture-of-Experts Quantization  
By Wanqi Yang, Yuexiao Ma, Alexander Conzelmann, Xiawu Zheng, Michael W. Mahoney, T. Konstantin Rusch, and Shiwei Liu  
Mixture-of-Experts (MoE) architectures scale computation via sparse expert activations, yet they remain memory-bound because all expert weights must reside in memory.

## Zero-Overhead Introspection for Adaptive Test-Time Compute  
By Rohin Manvi, Joey Hong, Tim Seyde, Maxime Labonne, Mathias Lechner, and Sergey Levine  
Large language models excel at reasoning but lack key aspects of introspection, including the ability to anticipate their own success and the computation required to achieve it.

## Learning residue level protein dynamics with multiscale Gaussians  
By Mihir Bafna, Bowen Jing, and Bonnie Berger  
Many methods have been developed to predict static protein structures, however understanding the dynamics of protein structure is essential for elucidating biological function.

## Antislop: A Comprehensive Framework for Identifying and Eliminating Repetitive Patterns in Language Models  
By Samuel J Paech, Allen G Roush, Judah Goldfeder, and Ravid Shwartz-Ziv  
Antislop detects and removes repetitive LLM “slop” using inference-time suppression, profiling, and targeted fine-tuning.

## The Curious Case of In-Training Compression of State Space Models  
By Makram Chahine, Philipp Nazari, Daniela Rus, and T. Konstantin Rusch  
State Space Models (SSMs), developed to tackle long sequence modeling tasks efficiently, offer both parallelizable training and fast inference.

## LFM2 Technical Report  
By Alexander Amini, Anna Banaszak, Harold Benoit, Arthur Böök, Tarek Dakhran, Song Duong, Alfred Eng, Fernando Fernandes, Marc Härkönen, Anne Harrington, Ramin Hasani, and 22 others  
We present LFM2, a family of Liquid Foundation Models designed for efficient on-device deployment and strong task capabilities.

## RL Squeezes, SFT Expands: A Comparative Study of Reasoning LLMs  
By Kohsei Matsutani, Shota Takashiro, Gouki Minegishi, Takeshi Kojima, Yusuke Iwasawa, and Yutaka Matsuo  
Large language models (LLMs) are typically trained by reinforcement learning (RL) with verifiable rewards and supervised fine-tuning (SFT) on reasoning traces to improve their reasoning abilities.

## Exploring Diffusion Transformer Designs via Grafting  
By Keshigeyan Chandrasegaran, Michael Poli, Daniel Y. Fu, Dongjun Kim, Lea M. Hadzic, Manling Li, Agrim Gupta, Stefano Massaroli, Azalia Mirhoseini, Juan Carlos Niebles, Stefano Ermon, and 1 other  
Grafting edits pretrained diffusion transformers into efficient hybrids using <2% compute, delivering strong quality and up to 1.43x speedups.

## Quantifying Memory Utilization with Effective State-Size  
By Rom N. Parnichkun, Neehal Tumma, Armin W. Thomas, Alessandro Moro, Qi An, Taiji Suzuki, Atsushi Yamashita, Michael Poli, and Stefano Massaroli  
Introducing effective state-size (ESS), a metric that quantifies memory utilization in sequence models to improve initialization and distillation.

## STAR: Synthesis of Tailored Architectures  
By Armin W. Thomas, Rom Parnichkun, Alexander Amini, Stefano Massaroli, and Michael Poli  
STAR uses evolutionary search over architecture genomes to automatically design hybrids that beat Transformers on quality, size, and cache.

## Towards a theory of learning dynamics in deep state space models  
By Jakub Smekal, Jimmy Smith, Michael Kleinman, Dan Biderman, and Scott Linderman  
Towards a theory of learning dynamics in deep state space models.

## The Empirical Impact of Neural Parameter Symmetries, or Lack Thereof  
By Derek Lim, Theo Putterman, Robin Walters, Haggai Maron, and Stefanie Jegelka  
The Empirical Impact of Neural Parameter Symmetries, or Lack Thereof.

## Large Scale Dataset Distillation with Domain Shift  
By Noel Loo, Alaa Maalouf, Ramin Hasani, Mathias Lechner, Alexander Amini, and Daniela Rus  
Large Scale Dataset Distillation with Domain Shift.

## LLM and Simulation as Bilevel Optimizers: A New Paradigm to Advance Physical Scientific Discovery  
By Pingchuan Ma, Tsun-Hsuan Wang, Minghao Guo, Zhiqing Sun, Joshua B. Tenenbaum, Daniela Rus, Chuang Gan, and Wojciech Matusik  
LLM and Simulation as Bilevel Optimizers: A New Paradigm to Advance Physical Scientific Discovery.

## RoboGen: Towards Unleashing Infinite Data for Automated Robot Learning via Generative Simulation  
By Yufei Wang, Zhou Xian, Feng Chen, Tsun-Hsuan Wang, Yian Wang, Katerina Fragkiadaki, Zackory Erickson, David Held, and Chuang Gan  
RoboGen: Towards Unleashing Infinite Data for Automated Robot Learning via Generative Simulation.

## Position: Future Directions in the Theory of Graph Machine Learning  
By Christopher Morris, Fabrizio Frasca, Nadav Dym, Haggai Maron, Ismail Ilkan Ceylan, Ron Levie, Derek Lim, Michael M. Bronstein, Martin Grohe, and Stefanie Jegelka  
Position: Future Directions in the Theory of Graph Machine Learning.

## Mechanistic Design and Scaling of Hybrid Architectures  
By Michael Poli, Armin W Thomas, Eric Nguyen, Pragaash Ponnusamy, Björn Deiseroth, Kristian Kersting, Taiji Suzuki, Brian Hie, Stefano Ermon, Christopher Re, Ce Zhang, and 1 other  
Mechanistic Design and Scaling of Hybrid Architectures.

## State-Free Inference of State-Space Models: The *Transfer Function* Approach  
By Rom Parnichkun, Stefano Massaroli, Alessandro Moro, Jimmy T.H. Smith, Ramin Hasani, Mathias Lechner, Qi An, Christopher Re, Hajime Asama, Stefano Ermon, Taiji Suzuki, and 2 others  
State-Free Inference of State-Space Models: The *Transfer Function* Approach.

## Growing Q-Networks: Solving Continuous Control Tasks with Adaptive Control Resolution  
By Tim Seyde, Peter Werner, Wilko Schwarting, Markus Wulfmeier, and Daniela Rus  
Growing Q-Networks: Solving Continuous Control Tasks with Adaptive Control Resolution.

## Understanding Reconstruction Attacks with the Neural Tangent Kernel and Dataset Distillation  
By Noel Loo, Ramin Hasani, Mathias Lechner, Alexander Amini, and Daniela Rus  
Understanding Reconstruction Attacks with the Neural Tangent Kernel and Dataset Distillation.

## Graph Metanetworks for Processing Diverse Neural Architectures  
By Derek Lim, Haggai Maron, Marc T. Law, Jonathan Lorraine, and James Lucas  
Graph Metanetworks for Processing Diverse Neural Architectures.