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

etric

-0.72

imedia

-0.71

vertisements

-0.70

itative

-0.69

otos

-0.69

ship

-0.69

ital

-0.69

owners

-0.68

trade

-0.68

mentioned

-0.67

POSITIVE LOGITS

bourg

0.80

ĸļ

0.78

URES

0.75

 Aires

0.73

 rand

0.72

gow

0.72

ĨĴ

0.71

pard

0.67

bian

0.66

gian

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.