INDEX

Explanations

references to various forms of powder

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 12-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_12/width_65k/average_l0_72

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 calendriers

-1.03

=<?

-0.99

BeginContext

-0.92

 للاسماء

-0.88

 Soria

-0.88

thmus

-0.87

 Greenway

-0.87

躇

-0.85

 DiCaprio

-0.85

thâu

-0.83

POSITIVE LOGITS

 Powder

1.63

 powder

1.52

Powder

1.51

 powders

1.38

powder

1.34

POW

1.13

Pow

1.12

 powdered

1.09

 powdery

1.03

POW

0.99

Activations Density 0.009%

references to various forms of powder

No Comments

No Known Activations