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

 Truman

-0.73

hood

-0.71

ston

-0.67

 Miracle

-0.62

yt

-0.62

%.

-0.62

dale

-0.60

aya

-0.60

guard

-0.59

Saf

-0.58

POSITIVE LOGITS

 guiActiveUnfocused

0.80

tnc

0.78

 concess

0.75

FTWARE

0.74

rent

0.73

aukee

0.73

ospace

0.73

)=(

0.72

PDATE

0.72

 partName

0.71

Activations Density 0.000%

No Known Activations

This feature has no known activations.