INDEX

Explanations

the term "Sa" and its variations, indicating references to a specific subject or concept

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 39-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_39/width_16k/average_l0_131

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.39.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

 addSubview

-0.43

dAtA

-0.40

lesias

-0.39

 Tara

-0.38

bital

-0.38

Elmer

-0.37

 Kindes

-0.36

book

-0.36

Tara

-0.36

CDCl

-0.36

POSITIVE LOGITS

httphttps

0.48

 disambiguazione

0.43

хьтан

0.42

 Scalars

0.41

 pamamagitan

0.40

 gonzález

0.39

 gawas

0.39

 myſelf

0.39

exitRule

0.38

profits

0.38

Activations Density 0.216%

the term "Sa" and its variations, indicating references to a specific subject or concept

No Comments

No Known Activations