INDEX

Explanations

references to contributors or authorship

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

copyright and contributors notice

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

ConstraintMaker

-0.60

ModelForm

-0.60

OrBuilder

-0.46

nationality

-0.40

scrapy

-0.38

 immagini

-0.38

Long

-0.32

Hashtable

-0.31

 judul

-0.31

Abt

-0.31

POSITIVE LOGITS

createCanvas

2.84

velocityX

1.54

velocityY

1.44

RemoveField

1.30

 createSprite

1.27

ModelAdmin

0.89

 CONTRIBUTORS

0.85

onViewCreated

0.79

StructEnd

0.72

featureID

0.71

Activations Density 0.000%

references to contributors or authorship

copyright and contributors notice

No Comments

No Known Activations

references to contributors or authorship

copyright and contributors notice

No Comments

No Known Activations