INDEX

Explanations

references to programming constructs or code snippets

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

Test types

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

invokeLater

-0.49

 ✭✭

-0.44

WritableDatabase

-0.42

 للاسماء

-0.40

nezeu

-0.40

 resourceCulture

-0.39

fromnode

-0.39

__(/*!

-0.38

UseVisualStyle

-0.38

 trampa

-0.37

POSITIVE LOGITS

 Савезне

0.67

 intStringLen

0.55

ViewImports

0.54

цездатний

0.54

 الرياضيه

0.46

 Infórmanos

0.46

 himo

0.45

 protoimpl

0.45

 fieldNum

0.44

 الاطلاع

0.42

Activations Density 0.039%

references to programming constructs or code snippets

Test types

No Comments

No Known Activations