INDEX

Explanations

Code and names

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_35.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.35.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

foil

-0.29

æĬľ

-0.28

 Silent

-0.28

æĹ©å°±

-0.27

tlement

-0.27

 Prel

-0.26

æŀľæĸŃ

-0.26

 Plastic

-0.25

æ¼ı

-0.25

ÐŁÐ¾Ð´ÑĢÐ¾Ð±

-0.25

POSITIVE LOGITS

anch

0.37

anche

0.35

ank

0.34

anko

0.32

æ°ĳæĹıæĸĩåĮĸ

0.32

anches

0.31

anka

0.30

enger

0.30

ate

0.30

aki

0.30

Activations Density 0.012%

Code and names

No Known Activations