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

 Frey

-0.82

ĸļ

-0.75

))))

-0.68

hirt

-0.68

gee

-0.67

opy

-0.67

ļéĨĴ

-0.67

fully

-0.66

ESE

-0.66

FAR

-0.63

POSITIVE LOGITS

utions

0.90

inances

0.87

eworthy

0.82

 prohibitions

0.69

 withdrawals

0.66

inctions

0.65

rences

0.65

 massacres

0.63

ineries

0.63

Pred

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.