llmfan46/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved-GGUF
Image-Text-to-Text • Updated • 70.4k • 71
Facu Vlad J
groxaxo
AI & ML interests
Cloud Engineer · AI Automation Engineer · Quantization Gremlin
I build cloud systems, automate the boring parts, and squeeze absurd efficiency out of AI models. Into infra, agents, vLLM, local GPU rigs, and quantizations that make big models run where they probably shouldn’t.
Recent Activity
liked a model about 6 hours ago
llmfan46/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved-GGUF liked a model about 6 hours ago
theoracleguy/Chatterbox-Multilingual-MLX-v2-Q8 liked a model about 22 hours ago
neuphonic/neutts-nano-spanish-q4-ggufOrganizations
None yet
liked 2
models about 6 hours ago
liked 2
models about 22 hours ago
reacted to RiverRider's post with 🔥 1 day ago
Post
4695
SRT-introspect: Live Token-by-Token Readout of LLM Internal Reasoning
I have released SRT-introspect, a new public demonstration that makes the hidden reasoning process of a frozen large language model visible in real time.
The interface runs a Qwen-2.5-7B backbone equipped with the SRT Adapter and Activation Verbalizer. As the model generates each token, the system continuously measures divergence across attention heads, identifies high-signal moments, and translates the corresponding hidden-state object representations into natural-language verbalizations. You see exactly what the model is internally representing at the precise points where its computation is most active, complete with divergence scores, reflexivity estimates, and per-layer traces.
This is not a summary of the final output. It is a direct window into the model’s latent conceptual landscape, showing the dominant training-data attractors that activate even when the prompt asks for first-principles reasoning. The adaptive scheduler concentrates verbalizations precisely where the real internal work occurs, turning what used to be opaque black-box generation into observable, analyzable data.
The result is the clearest public demonstration yet that modern LLMs possess a rich, structured semiotic infrastructure that can now be audited without retraining or fine-tuning.
Try it:
RiverRider/srt-introspect
I have released SRT-introspect, a new public demonstration that makes the hidden reasoning process of a frozen large language model visible in real time.
The interface runs a Qwen-2.5-7B backbone equipped with the SRT Adapter and Activation Verbalizer. As the model generates each token, the system continuously measures divergence across attention heads, identifies high-signal moments, and translates the corresponding hidden-state object representations into natural-language verbalizations. You see exactly what the model is internally representing at the precise points where its computation is most active, complete with divergence scores, reflexivity estimates, and per-layer traces.
This is not a summary of the final output. It is a direct window into the model’s latent conceptual landscape, showing the dominant training-data attractors that activate even when the prompt asks for first-principles reasoning. The adaptive scheduler concentrates verbalizations precisely where the real internal work occurs, turning what used to be opaque black-box generation into observable, analyzable data.
The result is the clearest public demonstration yet that modern LLMs possess a rich, structured semiotic infrastructure that can now be audited without retraining or fine-tuning.
Try it:
RiverRider/srt-introspect
liked 9
models 2 days ago
k2-fsa/sherpa-onnx-models
Updated • 2
FINAL-Bench/Darwin-60B-DUO
Text Generation • Updated • 398 • 29
FluidInference/pocket-tts-coreml
Text-to-Speech • Updated • 2.97k • 6
crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP
27B • Updated • 1.83k • 3
litmudoc/Chatterbox-Multilingual-MLX-v2-Q4
Text-to-Speech • 0.3B • Updated • 27 • 2
huginnfork/Qwen3.6-27B-uncensored-heretic-v2-mtp-FP8
Image-Text-to-Text • 28B • Updated • 2.75k • 2
drawais/Qwen3-Embedding-0.6B-AWQ-INT4
Text Generation • 0.6B • Updated • 3.23k • 3
Qwen/Qwen3-Embedding-0.6B
Feature Extraction • 0.6B • Updated • 8.41M • • 1.04k
Reza2kn/LocateAnything-3B-ONNX-WebGPU-INT4
Object Detection • Updated • 106 • 2
liked 5
models 3 days ago
FluidInference/qwen3-asr-0.6b-coreml
Automatic Speech Recognition • Updated • 238 • 11
FluidInference/speaker-diarization-coreml
Voice Activity Detection • Updated • 17.4k • 38
FluidInference/qwen3-tts-coreml
Text-to-Speech • Updated • 48 • 2
FluidInference/silero-vad-coreml
Voice Activity Detection • Updated • 28.7k • 24
FluidInference/parakeet-tdt-0.6b-v3-coreml
Automatic Speech Recognition • Updated • 254k • 45
liked a
model 4 days ago
