Hugging Face
Models
Datasets
Spaces
Buckets
new
Docs
Enterprise
Pricing
Website
Tasks
HuggingChat
Collections
Languages
Organizations
Community
Blog
Posts
Daily Papers
Hardware
Learn
Discord
Forum
GitHub
Solutions
Team & Enterprise
Hugging Face PRO
Enterprise Support
Inference Providers
Inference Endpoints
Storage Buckets
Log In
Sign Up
🌌
Is AI conscious?
65.0
TFLOPS
Banaxi
PRO
Banaxi-Tech
170
16
143
Follow
Shawnuc's profile picture
edenadiv's profile picture
ceaderf's profile picture
191 followers
·
74 following
Banaxi-Tech
AI & ML interests
SLMs, training from scratch, LoRA, TTS, Ternary models. AI Interpretability. BCI. Contact at banaxitech@gmail.com
Recent Activity
new
activity
about 7 hours ago
Qwen/Qwen3.8-2.4T-A95B:
Qwen3.8 16B
new
activity
about 13 hours ago
BananaMind/BananaMind-2.1-Unified:
Space for using this in HF?
replied
to
their
post
about 17 hours ago
We're releasing BananaMind 2.1 Unified, a 35M three-tower relay model where the two output towers can only talk to each other through a silent middle tower that has no output head and no loss term. Tower B trains entirely on indirect gradient. It was never told what to predict. It woke up anyway. Jacobian lens shows it carrying the correct answer ("Paris", "oxygen", "blue") at its deepest layer. Its bridge gates grew 4-47x from init. Feed it from only one side and the representations collapse to junk — it needs both outer towers to become semantic. The PIQA result is the cleanest demonstration: Tower A alone scores 50.11 (chance). Tower C alone 52.12. Full system 61.75. All physical reasoning lives in the integration. The 35M three-tower beats the 50M single-tower BananaMind 2 Medium on PIQA. Trained on 38B tokens in ~7.5 hours on 8x RTX PRO 6000. Ships with 7 ablation modes so you can surgically cut the model apart without retraining. Full training logs, J-lens fits, and eval outputs for every mode included. This is a research model. It is interesting because you can take it apart. https://huggingface.co/BananaMind/BananaMind-2.1-Unified Follow us for more: https://huggingface.co/BananaMind @Banaxi-Tech
View all activity
Organizations
Banaxi-Tech
's models
25
Sort: Recently updated
Banaxi-Tech/unified-2.1-test
Text Generation
•
35M
•
Updated
2 days ago
•
1.23k
Banaxi-Tech/pico-5
Text Generation
•
937k
•
Updated
8 days ago
•
176
Banaxi-Tech/pico-20
Text Generation
•
937k
•
Updated
8 days ago
•
178
Banaxi-Tech/pico-30
Text Generation
•
937k
•
Updated
8 days ago
•
170
Banaxi-Tech/pico-50
Text Generation
•
937k
•
Updated
8 days ago
•
165
Banaxi-Tech/pico-60
Text Generation
•
937k
•
Updated
8 days ago
•
164
Banaxi-Tech/pico-70
Text Generation
•
937k
•
Updated
8 days ago
•
164
Banaxi-Tech/pico-80
Text Generation
•
937k
•
Updated
8 days ago
•
107
Banaxi-Tech/test-pico-chinchilla
Text Generation
•
937k
•
Updated
8 days ago
•
130
Banaxi-Tech/pico-test
Text Generation
•
937k
•
Updated
8 days ago
•
777
Banaxi-Tech/slmoe-test
Text Generation
•
27.5M
•
Updated
9 days ago
•
779
Banaxi-Tech/custom-optimizer-model-test
Text Generation
•
12.1M
•
Updated
10 days ago
•
141
Banaxi-Tech/muon-model-test
Text Generation
•
12.1M
•
Updated
10 days ago
•
126
Banaxi-Tech/adamw-model-test
Text Generation
•
12.1M
•
Updated
10 days ago
•
145
Banaxi-Tech/instructions-for-challenge
Updated
11 days ago
Banaxi-Tech/BananaMind-FaceDetect-V1
Object Detection
•
2.24M
•
Updated
Jun 18
•
7
•
2
Banaxi-Tech/BananaMind-V2.5-Content-Safety
Text Generation
•
5B
•
Updated
Jun 11
•
102
•
2
Banaxi-Tech/BananaMind-Completor-V1
Text Generation
•
1.93M
•
Updated
Jun 4
•
133
•
2
Banaxi-Tech/BananaMind-TTS-V2
Text-to-Speech
•
9.5M
•
Updated
Jun 4
•
43
•
4
Banaxi-Tech/BananaMind-TTS-V2.1-Preview
Text-to-Speech
•
Updated
Jun 3
•
56
•
3
Banaxi-Tech/BananaMind-TTS-V1
Text-to-Speech
•
9.5M
•
Updated
May 26
•
6
•
1
Banaxi-Tech/BananaMind-Translate-V3.5-Pro
Translation
•
8B
•
Updated
May 24
•
10
•
3
Banaxi-Tech/BananaMind-Translate-V3.5
Translation
•
5B
•
Updated
May 24
•
12
•
1
Banaxi-Tech/BananaMind-Translate-V3.4
Translation
•
5B
•
Updated
May 11
•
29
•
2
Banaxi-Tech/BananaMind-v1
0.1B
•
Updated
Apr 20
•
20