INDEX

Explanations

dead

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_27.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.27.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

 promot

-0.26

 toJSON

-0.25

 promotion

-0.25

 Consultants

-0.25

HeaderText

-0.24

 consultant

-0.24

 promotions

-0.24

 little

-0.24

 mentality

-0.23

little

-0.23

POSITIVE LOGITS

bras

0.28

Ø¨Ø©

0.27

emplate

0.26

assing

0.26

 kend

0.25

éĩįå¤į

0.24

ĲèĹı

0.24

forward

0.24

æŃ¦

0.24

-center

0.24

Activations Density 0.008%

dead

No Comments

No Known Activations