INDEX

Explanations

topp

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_30.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.30.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

ÑĢÑĥÐ´

-0.27

çŃĴ

-0.27

andal

-0.26

å©Ĩ

-0.25

AGING

-0.25

_USED

-0.25

ITAL

-0.24

itech

-0.24

 magazines

-0.23

hus

-0.23

POSITIVE LOGITS

olicies

0.26

arser

0.26

ä½ĵèĤ²åľº

0.26

area

0.25

olle

0.25

arro

0.25

çĬ¶

0.25

å¾ĹæĦı

0.25

çļĦåĬ¨åĬĽ

0.25

åį¡æĭī

0.25

Activations Density 0.001%

topp

No Known Activations