INDEX

Explanations

on the one

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_31.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.31.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 handle

-0.50

handle

-0.49

 handling

-0.48

 handles

-0.45

 Handling

-0.44

_handle

-0.43

 handled

-0.42

Handling

-0.42

 Handle

-0.41

Handle

-0.41

POSITIVE LOGITS

æĮĩ

0.32

=end

0.28

 endDate

0.28

 fingers

0.27

flate

0.26

berman

0.26

jured

0.26

åľºæ¯ĶèµĽ

0.26

ç«¯

0.26

 Ende

0.25

Activations Density 0.105%

on the one

No Known Activations