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

iott

-0.76

 Flavoring

-0.76

iffin

-0.73

 Amos

-0.69

opsy

-0.65

inez

-0.65

aido

-0.63

rex

-0.63

 Walls

-0.63

 Tiff

-0.63

POSITIVE LOGITS

sal

1.27

 recl

0.73

 spoiled

0.69

oct

0.69

 incl

0.67

Quantity

0.66

 glare

0.66

 mould

0.66

 acre

0.65

bay

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.