© 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 · 18535 ｜ Neuronpedia

Home
Qwen3-32B
32-RESID-BATCHTOPK-65K
18535

INDEX

Explanations

lab contexts

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

ality

-0.13

nable

-0.10

lish

-0.10

DBG

-0.09

ly

-0.09

ances

-0.09

lint

-0.09

chemes

-0.08

ably

-0.08

åľ

-0.08

POSITIVE LOGITS

rador

0.17

elling

0.12

NewLabel

0.10

rats

0.10

mates

0.10

/lab

0.10

rus

0.10

é©·

0.10

å°½å¤´

0.10

rique

0.09

Activations Density 0.027%

No Known Activations