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

agra

-0.77

rolet

-0.68

committee

-0.68

ugu

-0.67

Leaks

-0.67

SpaceEngineers

-0.66

Thumbnail

-0.66

prus

-0.66

pmwiki

-0.65

soDeliveryDate

-0.64

POSITIVE LOGITS

inates

0.70

boarding

0.66

 Grind

0.63

++++++++++++++++

0.62

 preferred

0.60

 Cyan

0.59

Santa

0.59

 Dove

0.59

Gray

0.59

Hal

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.