© 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 · 16998 ｜ Neuronpedia

Home
Qwen3-32B
32-RESID-BATCHTOPK-65K
16998

INDEX

Explanations

against

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

æĬĬæīĭ

-0.09

ihat

-0.09

IFT

-0.09

å¤´ä¸Ĭ

-0.08

æĭ¿äºĨ

-0.08

å±¥

-0.08

zew

-0.08

ç¿°

-0.08

æĬ¢éĻ©

-0.08

æĬĺ

-0.08

POSITIVE LOGITS

 against

0.13

against

0.13

å¢ĻéĿ¢

0.13

Against

0.13

 tÆ°á»Ŀng

0.12

å¢Ļå£ģ

0.12

 wall

0.11

å¤ĸå¢Ļ

0.11

 Against

0.11

 contre

0.10

Activations Density 0.087%

No Known Activations