INDEX

Explanations

No Explanations Found

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_20.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.20.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 safest

-0.29

æľºæ¢°

-0.28

otech

-0.27

ãģĶ

-0.25

 safety

-0.25

 mechanical

-0.25

éĺľ

-0.24

 SHALL

-0.24

ãĥĲãĥ©ãĥ³ãĤ¹

-0.24

åĭį

-0.24

POSITIVE LOGITS

è§£æĶ¾æĢĿæĥ³

0.30

Sof

0.29

reve

0.28

 wireType

0.26

ç©¿éĢı

0.26

PLACE

0.26

 rebel

0.25

æŁĲæŁĲ

0.25

åĩºæĿ¥

0.25

è§ģè¯ģäºĨ

0.24

Activations Density 2.540%

No Known Activations

This feature has no known activations.