INDEX

Explanations

preprocessor directives in code, specifically 'define' statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-mlp-65k

Configuration

google/gemma-scope-2b-pt-mlp/layer_0/width_65k/average_l0_72

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.0.hook_mlp_out

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 define

-1.02

define

-0.96

Define

-0.80

 defines

-0.73

 ExecuteAsync

-0.71

 Exactos

-0.69

defines

-0.69

 Define

-0.69

definition

-0.68

 defined

-0.66

POSITIVE LOGITS

PhysRevD

0.83

IBOutlet

0.67

 Hampden

0.66

 yağı

0.63

 grunn

0.62

lawayo

0.62

Pune

0.61

 Marathi

0.61

 Anfitrión

0.60

 Zink

0.60

Activations Density 0.025%

preprocessor directives in code, specifically 'define' statements

No Comments

No Known Activations