INDEX

Explanations

geographical location names and their associated regions

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

Not in Any Lists

No Comments

Negative Logits

 enfans

-0.95

 pleaſure

-0.80

ſelf

-0.78

 stanga

-0.72

ſelves

-0.72

 itſelf

-0.71

uș

-0.69

 myſelf

-0.69

 avoient

-0.68

 GenerationType

-0.68

POSITIVE LOGITS

ng

0.53

đ

0.51

kh

0.49

huy

0.47

th

0.47

thu

0.47

sau

0.46

nh

0.46

Đ

0.46

cho

0.45

Activations Density 0.029%

geographical location names and their associated regions

No Comments

No Known Activations