INDEX

Explanations

names of people or figures associated with various contexts

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Bo

-0.38

Bor

-0.36

Fra

-0.34

sha

-0.34

Re

-0.33

WebControls

-0.33

Fe

-0.33

Su

-0.33

Mar

-0.32

Si

-0.32

POSITIVE LOGITS

ंदीखरीदारी

0.79

 CreateTagHelper

0.79

IndentedString

0.76

 ब्रेकडाउन

0.75

AsUp

0.73

 незавершена

0.71

 Biôgrafia

0.70

 Verſ

0.68

 propOrder

0.66

 noDo

0.65

Activations Density 0.091%

names of people or figures associated with various contexts

No Comments

No Known Activations