INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

edom

-0.98

icut

-0.85

acca

-0.75

alian

-0.71

olulu

-0.70

uctions

-0.68

oral

-0.67

otos

-0.67

shaw

-0.67

eln

-0.66

POSITIVE LOGITS

bilt

0.75

ENDED

0.74

Ã£

0.71

Kab

0.70

 Shank

0.69

BMC

0.64

Kap

0.64

Bh

0.63

 BahÃ¡

0.63

vier

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.