INDEX

Explanations

various types of objects or devices, often emphasizing their functions or characteristics

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

 saree

-0.50

 hurricanes

-0.48

 Chains

-0.48

 earthquakes

-0.47

 Perfume

-0.47

liosis

-0.47

 shotguns

-0.47

 violins

-0.47

 katana

-0.47

 Rifles

-0.47

POSITIVE LOGITS

 objects

1.37

 items

1.13

 object

1.05

 benda

1.01

 objetos

1.00

 objets

1.00

 oggetti

0.98

objects

0.97

 item

0.94

アイテム

0.94

Activations Density 1.125%

various types of objects or devices, often emphasizing their functions or characteristics

No Comments

No Known Activations