INDEX

Explanations

phrases related to correspondence or relationships between concepts or elements

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-pt-res/layer_2/width_16k/average_l0_67

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.2.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

̈́

-0.57

 فريبيس

-0.56

 Contacts

-0.56

Vidite

-0.56

 ModelExpression

-0.55

Contact

-0.54

///</

-0.54

ellido

-0.53

rinfo

-0.53

 Contact

-0.52

POSITIVE LOGITS

 correspond

0.97

 corresponding

0.84

 corresponded

0.82

corresponding

0.82

 corresponds

0.82

 Corresponding

0.81

correspond

0.78

 Correspond

0.75

 corresponden

0.73

 corrispond

0.68

Activations Density 0.098%

phrases related to correspondence or relationships between concepts or elements

No Known Activations