INDEX

Explanations

variations of the term "sp" indicating a specific subject or entity within the text

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

endpush

-0.39

HostException

-0.33

om

-0.33

 PyTuple

-0.30

AnimationsModule

-0.30

Controllo

-0.30

labus

-0.29

 finns

-0.29

WriteHeader

-0.28

TextAppearance

-0.28

POSITIVE LOGITS

sp

0.81

Sp

0.69

 spiked

0.67

########.

0.60

sp

0.60

Sp

0.59

帖最后由

0.57

GEBURTSDATUM

0.56

 queſta

0.56

 Spraw

0.55

Activations Density 0.206%

variations of the term "sp" indicating a specific subject or entity within the text

No Known Activations