INDEX

Explanations

but

np_max-act · gemini-2.0-flash

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_12.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.12.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

éģĹæ¼ı

-0.25

thy

-0.25

å¾Ĥ

-0.25

 forgetting

-0.24

çļĦå§¿æĢģ

-0.24

ForMember

-0.24

 forget

-0.24

çİ¯å¢ĥä¸Ń

-0.24

 bleach

-0.23

 reopening

-0.23

POSITIVE LOGITS

uco

0.28

lique

0.28

uly

0.27

æ¥¼

0.27

ESCO

0.27

å°¾

0.26

ames

0.26

åıĪæĺ¯

0.26

ÑĥÑī

0.25

cup

0.25

Activations Density 0.002%

but

No Comments

No Known Activations