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

 Nicaragua

-0.75

EStreamFrame

-0.69

 Nicarag

-0.69

 Tarant

-0.68

 Ratt

-0.68

 Pengu

-0.67

anan

-0.65

 Twain

-0.65

ufact

-0.64

iences

-0.64

POSITIVE LOGITS

rad

0.87

sheet

0.82

erc

0.72

club

0.68

ded

0.68

ding

0.67

hed

0.67

ging

0.67

stretched

0.66

rock

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.