© Neuronpedia 2026

Privacy & Terms Blog GitHub Slack Twitter Contact

Neuronpedia

Jacobian LensNEW

Natural Language

NEW Assistant AxisNEW Circuit TracerUPDATESteer SAE Evals Exports interp-engineNEWAPI Community Blog Privacy & Terms Contact

QWEN3-32B · 32-RESID-BATCHTOPK-65K · 16982 ｜ Neuronpedia

Home
Qwen3-32B
32-RESID-BATCHTOPK-65K
16982

INDEX

Explanations

hum words

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Configuration

adamkarvonen/qwen3-32b-saes/saes_Qwen_Qwen3-32B_batch_top_k/resid_post_layer_32

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

No Configuration Found

Embeds

Show PlotsShow ExplanationShow ActivationsShow Test FieldShow SteerShow Link

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

bach

-0.11

hots

-0.09

ivas

-0.09

how

-0.08

owing

-0.08

WISE

-0.08

entr

-0.08

è¿ĽåĨĽ

-0.08

auce

-0.08

anders

-0.08

POSITIVE LOGITS

hum

0.13

Hum

0.12

Hum

0.11

è°¦

0.11

 humble

0.11

è®µ

0.10

hum

0.10

åįĳ

0.09

ghan

0.09

åĹŁ

0.09

Activations Density 0.004%

No Known Activations