INDEX

Explanations

technical and programming-related terms including data types and structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 27-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_27/width_16k/average_l0_118

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.27.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

agenda

-0.59

Nug

-0.57

cupa

-0.56

Mila

-0.56

ing

-0.56

 Esco

-0.55

 pellic

-0.54

Esco

-0.53

 Lark

-0.53

AddTagHelper

-0.53

POSITIVE LOGITS

 navideña

0.54

 fotográfica

0.50

 zpě

0.49

 sostenible

0.47

 pelatih

0.46

 penderita

0.45

 nocturno

0.45

 deportivo

0.45

 navideño

0.45

 chrétien

0.43

Activations Density 1.725%

technical and programming-related terms including data types and structures

No Comments

No Known Activations