INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

erb

-0.70

alys

-0.70

 tremend

-0.69

bench

-0.69

andro

-0.69

haar

-0.68

 Meal

-0.67

opal

-0.66

flo

-0.66

Redd

-0.64

POSITIVE LOGITS

 Presidents

0.74

 Governor

0.70

Southern

0.69

 Governors

0.68

STATE

0.66

cedented

0.66

 governor

0.65

 governors

0.63

AES

0.62

 Prohibition

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.