INDEX

Explanations

instances of objects being hidden or concealed within other objects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 37-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_37/width_16k/average_l0_124

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.37.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

 AssemblyCompany

-0.49

 stationed

-0.46

Stored

-0.45

 detained

-0.44

endor

-0.42

 stored

-0.42

 parked

-0.42

LayoutStyle

-0.42

 aufges

-0.41

 deployed

-0.41

POSITIVE LOGITS

 securely

0.83

 somewhere

0.76

 inside

0.73

 firmly

0.65

 atop

0.63

somewhere

0.63

 snug

0.63

 neatly

0.62

inside

0.62

 tightly

0.61

Activations Density 0.723%

instances of objects being hidden or concealed within other objects

No Comments

No Known Activations