INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ignt

-0.83

ĪĴ

-0.74

ettings

-0.74

Versions

-0.73

Cos

-0.72

Apps

-0.72

));

-0.71

hips

-0.71

ãĤ¢ãĥ«

-0.71

wings

-0.69

POSITIVE LOGITS

 ounce

0.72

ization

0.68

 phosph

0.68

 penetration

0.66

 moderator

0.62

ium

0.61

XL

0.61

 occupation

0.60

CFR

0.60

 mention

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.