Akhiad Bercovich

Akhiad Bercovich has 4 papers in our library, in 2025, most of them on Reasoning, Language models & pretraining; often with Abhinav Khattar, Aleksander Ficek, Aarti Basant.

4 papers in our library · 2025

The profile is the library's match of the name; two people with one name are not always told apart.

Fields, topics and concepts

Of its works in our library, with the number of works on each.

The most cited first (OpenAlex and our library).

Llama-Nemotron: Efficient Reasoning Models (2025 · 76 citations)
Reasoning-oriented language models write long chains of thought, so inference cost increasingly limits how useful they are.
NVIDIA Nemotron Nano 2: An Accurate and Efficient Hybrid Mamba-Transformer Reasoning Model (2025 · 39 citations)
Reasoning models spend most of their compute producing long chains of thought. In standard Transformers, the attention KV cache makes such long generations slow and memory-hungry. NVIDIA addresses this with Nemotron-Nano-9B-v2, which uses the Nemotron-H hybrid design: most self-attention layers are replaced by…
Nemotron-H: A Family of Accurate and Efficient Hybrid Mamba-Transformer Models (2025 · 38 citations)
Self-attention costs grow with sequence length, both in compute per generated token and in KV-cache memory.
Nemotron 3 Nano: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning (2025 · 27 citations)
NVIDIA presents Nemotron 3 Nano 30B-A3B, an open language model for agentic use, reasoning and chat that aims for a better balance between accuracy and inference speed.

5 works of theirs in our library, the most cited first; 4 with a page here, the others link their original until we summarise them.

10 co-authors with a page, the most papers together first.

Papers of our library cite their works 187 times (of the 149,439 papers whose bibliographies we read); 41 citing papers with a page here →.

Updated 30 October 2026