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

 Dream

-0.68

ultan

-0.68

kaya

-0.66

Â¢

-0.66

 airing

-0.65

Swe

-0.63

maid

-0.63

-0.62

warming

-0.61

Ky

-0.61

POSITIVE LOGITS

 behavi

0.76

oters

0.73

ONES

0.72

STAT

0.71

theless

0.69

iban

0.69

ailability

0.68

igham

0.67

olean

0.67

Phys

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.