INDEX

Explanations

No Explanations Found

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_20.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.20.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

orus

-0.29

ä»Ģä¹Īåĳ¢

-0.27

udes

-0.27

éĿ¢èĨľ

-0.26

 nulla

-0.25

æĭį

-0.25

atorium

-0.24

è¶ĭ

-0.24

Ð¸Ð¼Ð²

-0.24

UnderTest

-0.23

POSITIVE LOGITS

 Frag

0.28

 together

0.28

æ

0.26

çªĸ

0.26

 Together

0.26

 pockets

0.25

 frag

0.25

mo

0.24

oter

0.24

 radio

0.24

Activations Density 2.659%

No Known Activations

This feature has no known activations.