INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Himself

-0.68

 Helpful

-0.66

orsi

-0.64

bite

-0.63

 numb

-0.62

PsyNetMessage

-0.62

 allerg

-0.62

Exception

-0.61

ILY

-0.61

Cth

-0.61

POSITIVE LOGITS

interstitial

0.89

bilt

0.71

ACTED

0.67

-->

0.64

Bow

0.63

 Belt

0.63

 fiber

0.63

actionDate

0.62

 eyed

0.61

uilding

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.