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

 Phillips

-0.72

 Mons

-0.70

Rez

-0.70

Phill

-0.70

 Dover

-0.69

Ps

-0.69

 Fast

-0.68

 PARK

-0.67

 Paddock

-0.67

 Desk

-0.66

POSITIVE LOGITS

alty

0.93

alien

0.82

peria

0.81

 indebted

0.80

 subtitle

0.75

ebted

0.74

ervative

0.73

awaru

0.73

 addicted

0.72

 interest

0.70

Activations Density 0.000%

No Known Activations

This feature has no known activations.