INDEX

Explanations

that

np_max-act · gemini-2.0-flash

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_23.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.23.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

boy

-0.29

chied

-0.27

æľīäºĽ

-0.25

ey

-0.25

 sometimes

-0.25

ry

-0.24

 exceptions

-0.24

ä¾ĭå¤ĸ

-0.24

<object

-0.23

è§Ĵèī²

-0.23

POSITIVE LOGITS

eneration

0.30

 contracting

0.28

tout

0.27

æľīæĦı

0.26

 contraction

0.26

 Charging

0.26

_correction

0.25

Charge

0.25

 Extraction

0.24

 elimination

0.24

Activations Density 0.003%

that

No Comments

No Known Activations