INDEX

Explanations

terms related to embedded structures or methods

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 2-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_2/width_16k/average_l0_67

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.2.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

 GenerationType

-0.63

 otomatig

-0.61

MigrationBuilder

-0.53

hésite

-0.50

exitRule

-0.47

ScopeManager

-0.47

 CommandType

-0.45

 bookId

-0.45

EndContext

-0.45

VersionUID

-0.45

POSITIVE LOGITS

 embedded

0.80

embed

0.73

 embed

0.71

Embed

0.69

embedded

0.69

 embedding

0.68

Embedding

0.66

 Embedding

0.63

 overwhelming

0.62

embedding

0.62

Activations Density 0.283%

terms related to embedded structures or methods

No Comments

No Known Activations