INDEX

Explanations

references to various items and their characteristics

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

Not in Any Lists

No Comments

Negative Logits

retudo

-0.39

 frete

-0.36

φέρ

-0.35

 Confira

-0.35

 interrump

-0.35

 motores

-0.35

 Priester

-0.33

 Kowalski

-0.33

 Wieder

-0.33

 cadeiras

-0.33

POSITIVE LOGITS

 something

0.66

小

0.65

 BoxFit

0.65

大

0.59

Personensuche

0.58

 item

0.58

 itſelf

0.57

NewUrlParser

0.54

something

0.52

 HasFactory

0.52

Activations Density 0.166%

references to various items and their characteristics

No Comments

No Known Activations