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

ulhu

-0.74

];

-0.73

 mathemat

-0.65

 totality

-0.65

 compromises

-0.63

 vector

-0.62

 Sorceress

-0.62

 directional

-0.62

 Grimm

-0.59

 vectors

-0.59

POSITIVE LOGITS

¿

0.75

ãĥ´

0.74

Ĩ

0.71

cia

0.69

ľ

0.68

Ĺ

0.68

lines

0.68

ãĥ¼ãĥ

0.67

rod

0.66

fur

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.