INDEX

Explanations

Software/code related

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

Not in Any Lists

No Comments

Negative Logits

ä»½

-0.27

/testify

-0.27

[('

-0.27

åĲĽ

-0.26

æ³¡

-0.26

æĹ§

-0.25

åħ³éĶ®

-0.25

perform

-0.25

åĸĦ

-0.25

.ico

-0.24

POSITIVE LOGITS

é¥´

0.27

fbe

0.27

 Viet

0.27

imson

0.27

anine

0.26

hr

0.26

æĪĺçķ¥å¸ĥå±Ģ

0.26

itel

0.24

 depressing

0.24

amins

0.24

Activations Density 0.018%

Software/code related

No Comments

No Known Activations