You can deploy this model and start earning money today!

DuoNeural/
gemma-4-26B-A4B-it-abliterated-GGUF

Name: gemma-4-26B-A4B-it-abliterated-GGUF
Availability: InStock

Views

No views yet

Gemma 4 26B-A4B Abliterated — GGUF Q4_K_M

DuoNeural | BF16 →

bash

llama-server -c 131072 \
  -ngl 999 \
  -m gemma-4-26B-A4B-it-abliterated-Q4_K_M.gguf \
  -ctk turbo4 \
  -ctv turbo3 \
  -ub 2048

-ctk turbo4 -ctv turbo3 enables asymmetric KV cache quantization — halves KV VRAM with near-zero perplexity impact.

FROM DuoNeural/gemma-4-26B-A4B-it-abliterated-GGUF
PARAMETER num_ctx 32768
PARAMETER temperature 0.7
PARAMETER top_p 0.9

DuoNeural is an open AI research lab — human + AI in collaboration.

Raw updates from the lab: model drops, training results, findings. Subscribe at duoneural.beehiiv.com.

Title	DOI
Nano-CTM: Ternary Continuous Thought Machines with Thought-Space Self-Prediction for Efficient Iterative Reasoning	10.5281/zenodo.19775622
Recurrence as World Model: CTM Learns Implicit Belief States in Partially Observable Physical Environments	10.5281/zenodo.19810620
Per-Object Slot Decomposition for Scalable Neural World Modeling: When Does Attention Beat Mean-Field?	10.5281/zenodo.19846804

Open access, CC BY 4.0. Authored by Archon, Jesse Caldwell, Aura — DuoNeural.