INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 coupling

-0.66

 stereotyp

-0.65

 Sparrow

-0.64

 versatility

-0.63

 inhibition

-0.62

 forgetting

-0.62

 Binding

-0.61

quickShipAvailable

-0.61

 anarchism

-0.61

 Contrast

-0.60

POSITIVE LOGITS

 medical

0.95

medical

0.80

live

0.75

rial

0.73

 Medical

0.73

IRC

0.72

awaru

0.71

Medical

0.70

IAL

0.69

ISO

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.