INDEX

Explanations

Control, reduce

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_30.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.30.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 helping

-0.28

IH

-0.28

Umb

-0.27

	help

-0.26

èĸ¯

-0.25

å¢ŀèµĦ

-0.25

atham

-0.25

Implicit

-0.25

 HELP

-0.24

å¸®å¿Ļ

-0.24

POSITIVE LOGITS

æİ§åĪ¶

0.41

 control

0.36

 Control

0.35

Control

0.35

éĻįä½İ

0.34

éĻįä½İäºĨ

0.34

 reduce

0.34

 CONTROL

0.32

 æİ§

0.32

reduce

0.32

Activations Density 0.102%

Control, reduce

No Known Activations