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

 Marcos

-0.74

 commissions

-0.70

achev

-0.70

 Ferdinand

-0.70

../

-0.68

 Paulo

-0.66

 Nieto

-0.64

 allocations

-0.62

stration

-0.62

imental

-0.62

POSITIVE LOGITS

áµ

0.83

Thor

0.75

Virginia

0.75

Dex

0.73

Reviewer

0.72

Amazing

0.72

Page

0.70

Fort

0.70

OX

0.70

arrow

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.