INDEX

Explanations

phrases conveying mistakes or errors

oai_token-act-pair · gpt-4o-mini Triggered by @bot

forgot to mention

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Népesség

-0.44

 الرياضيه

-0.41

 tartalomajánló

-0.40

“:

-0.39

новништво

-0.38

 CURIAM

-0.36

ChildScrollView

-0.35

‚

-0.34

 balanceOf

-0.33

moveToFirst

-0.32

POSITIVE LOGITS

 geweest

0.53

Kuz

0.48

帖最后由

0.48

webElementXpaths

0.48

itsub

0.47

 Discretion

0.47

Forgot

0.47

vably

0.46

 mistakenly

0.46

 initComponents

0.46

Activations Density 0.025%

phrases conveying mistakes or errors

forgot to mention

No Comments

No Known Activations