INDEX

Explanations

No Explanations Found

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_18.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.18.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

TBD

-0.26

iname

-0.25

 oral

-0.25

 achie

-0.25

oplay

-0.24

aseline

-0.23

iminary

-0.23

Ã©t

-0.23

 Dudley

-0.23

ä¸ºä¸»

-0.23

POSITIVE LOGITS

 escap

0.28

 bother

0.27

esty

0.27

èĥ¸

0.26

RH

0.26

åŀ®

0.26

 escapes

0.25

uish

0.25

rophe

0.25

 trade

0.25

Activations Density 3.392%

No Known Activations

This feature has no known activations.