INDEX

Explanations

Parts of words

np_max-act · gemini-2.0-flash

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_7.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.7.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

java

-0.27

 IMDb

-0.27

xBA

-0.25

åİŁæłĩé¢ĺ

-0.25

 decor

-0.25

 java

-0.24

 frags

-0.24

 mods

-0.24

aseline

-0.24

dec

-0.24

POSITIVE LOGITS

cki

0.29

-expand

0.27

FETCH

0.26

 competitiveness

0.26

ultipart

0.26

 competitive

0.25

ERV

0.25

Ð¾ÑĩÐ½ÑĭÑħ

0.25

erce

0.25

isoft

0.24

Activations Density 0.002%

Parts of words

No Comments

No Known Activations