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

forth

-0.71

rote

-0.67

ieved

-0.66

odium

-0.64

raught

-0.63

itled

-0.63

orate

-0.63

amount

-0.62

 amended

-0.62

cerpt

-0.60

POSITIVE LOGITS

Gameplay

0.75

gencies

0.74

 Moonlight

0.66

interstitial

0.66

ËĪ

0.66

ggies

0.64

endez

0.63

akeru

0.63

Episode

0.62

ivalry

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.