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

prus

-0.77

icz

-0.77

berus

-0.70

 carriers

-0.68

coni

-0.65

icians

-0.65

uggest

-0.64

renches

-0.64

 tubes

-0.63

angler

-0.63

POSITIVE LOGITS

ĵĺ

0.82

IFIED

0.73

¬¼

0.71

 PowerPoint

0.69

©¶æ¥µ

0.68

Kiw

0.66

 Wonderland

0.66

Yao

0.66

¶æ

0.64

¡

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.