INDEX

Explanations

instances of programming or licensing notation, particularly in comments

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

code comment symbols

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

Personendaten

-0.47

FieldBuilder

-0.46

glBindBuffer

-0.43

FieldDescriptor

-0.42

 positioning

-0.41

medriver

-0.41

EndInit

-0.41

 Ferri

-0.41

stdc

-0.40

 Mitar

-0.40

POSITIVE LOGITS

 kasarigan

0.65

 ब्रेकडाउन

0.55

 raiſ

0.49

帖最后由

0.46

GEBURTSDATUM

0.45

 auroit

0.44

 appartiennent

0.44

 supérieures

0.41

 Италијани

0.41

 supérieurs

0.40

Activations Density 0.013%

instances of programming or licensing notation, particularly in comments

code comment symbols

No Comments

No Known Activations