INDEX

Explanations

code and websites

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_25.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.25.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

memberof

-0.27

 dashes

-0.27

dash

-0.26

ç»ıèĲ¥æ´»åĬ¨

-0.26

è®ºè¯ģ

-0.26

 dash

-0.25

æ²¹èĦĤ

-0.25

è£¹

-0.25

æĬķèµĦé¡¹çĽ®

-0.25

æ±Ĥå©ļ

-0.25

POSITIVE LOGITS

che

0.30

 starter

0.28

 collaborators

0.26

Neighbors

0.26

 Collabor

0.26

æ§İ

0.26

.codes

0.26

=ax

0.25

 collaborating

0.25

 cheat

0.25

Activations Density 0.991%

code and websites

No Known Activations