INDEX

Explanations

advice on what to avoid

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_27.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.27.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

Thunk

-0.28

æµħ

-0.26

æĬĢæľ¯åĴĮ

-0.26

_instructions

-0.24

è¦ģåľ¨

-0.23

å»ºè®®

-0.23

é¢ĺ

-0.23

äºİæŃ¤

-0.23

å·²æĪĲä¸º

-0.23

ä¼ļåľ¨

-0.23

POSITIVE LOGITS

 always

0.28

 entirely

0.26

çĶ¸

0.26

pot

0.26

uhn

0.26

 like

0.25

strlen

0.25

 avoiding

0.25

_matching

0.24

_like

0.24

Activations Density 0.783%

advice on what to avoid

No Comments

No Known Activations