10 papers · 1 filter
Power-Softmax: Towards Secure LLM Inference over Encrypted Data
Itamar Zimerman, Allon Adir, Ehud Aharoni +7
Modern cryptographic methods for implementing privacy-preserving LLMs such as \gls{HE} require the LLMs to have a polynomial form. Forming such a representation is challenging beca…
TensorLens: End-to-End Transformer Analysis via High-Order Attention Tensors
Ido Andrew Atad, Itamar Zimerman, Shahar Katz +1
Attention matrices are fundamental to transformer research, supporting a broad range of applications including interpretability, visualization, manipulation, and distillation. Yet,…
Efficient Decoding Methods for Language Models on Encrypted Data
Matan Avitan, Moran Baruch, Nir Drucker +2
Large language models (LLMs) power modern AI applications, but processing sensitive data on untrusted servers raises privacy concerns. Homomorphic encryption (HE) enables computati…
Differential Mamba
Nadav Schneider, Itamar Zimerman, Eliya Nachmani
Sequence models like Transformers and RNNs often overallocate attention to irrelevant context, leading to noisy intermediate representations. This degrades LLM capabilities by prom…
Overflow Prevention Enhances Long-Context Recurrent LLMs
Assaf Ben-Kish, Itamar Zimerman, M. Jehanzeb Mirza +4
A recent trend in LLMs is developing recurrent sub-quadratic models that improve long-context processing efficiency. We investigate leading large long-context models, focusing on h…
Overclocking LLM Reasoning: Monitoring and Controlling Thinking Path Lengths in LLMs
Roy Eisenstadt, Itamar Zimerman, Lior Wolf
Recently, techniques such as explicit structured reasoning have demonstrated strong test-time scaling behavior by enforcing a separation between the model's internal "thinking" pro…