INDEX

Explanations

instances of numerical data or quantifiable values

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 12-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_12/width_16k/average_l0_130

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.12.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

abank

-0.67

 Bobo

-0.66

 Pinocchio

-0.63

alao

-0.62

 Blok

-0.60

談社

-0.60

 NSCoder

-0.60

สือ

-0.59

 Cordero

-0.59

 blockade

-0.57

POSITIVE LOGITS

1.57

７

0.88

 Seven

0.88

 seven

0.85

 zeven

0.84

۷

0.82

 SEVEN

0.81

 Seventh

0.79

 seventh

0.79

 seventy

0.78

Activations Density 0.704%

instances of numerical data or quantifiable values

No Comments

No Known Activations