INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

OUR

-0.80

owship

-0.72

ACC

-0.69

IND

-0.68

WHERE

-0.68

ENC

-0.67

ITAL

-0.66

Ja

-0.65

ANCE

-0.64

CENT

-0.64

POSITIVE LOGITS

eki

0.72

 Newtown

0.71

 Miko

0.71

 Becky

0.65

ãĥ¯

0.65

ãĥŀ

0.63

zbek

0.63

roe

0.62

ãĤµ

0.62

 Genie

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.