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

ONSORED

-0.78

Þ

-0.77

©¶æ

-0.74

 Cosponsors

-0.72

 adolesc

-0.69

isan

-0.69

ô

-0.69

IAL

-0.67

hner

-0.65

 Hits

-0.64

POSITIVE LOGITS

jured

0.70

readable

0.67

memory

0.67

arth

0.66

force

0.65

 emptiness

0.64

illusion

0.64

vision

0.63

ItemTracker

0.63

 blockade

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.