INDEX

Explanations

specific patterns or categories in scientific or technical contexts

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

molten metal

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

AddTagHelper

-0.52

parsedMessage

-0.52

TagMode

-0.44

OGND

-0.43

 elimination

-0.43

setAccessible

-0.42

hetic

-0.42

 Glou

-0.42

Koz

-0.42

⤹

-0.42

POSITIVE LOGITS

 AttributeSet

0.67

ftagPool

0.54

ViewFeatures

0.52

AnimationsModule

0.50

AttributeSet

0.48

astéroïdes

0.44

 struktur

0.43

 ComVisible

0.42

 kandid

0.42

 autorytatywna

0.41

Activations Density 0.015%

specific patterns or categories in scientific or technical contexts

molten metal

No Comments

No Known Activations