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

onet

-0.70

obby

-0.67

UC

-0.65

tom

-0.64

UTC

-0.63

noon

-0.62

BILITIES

-0.62

leneck

-0.62

EW

-0.61

externalActionCode

-0.61

POSITIVE LOGITS

interstitial

0.84

 Compass

0.71

 aven

0.70

ĸļ

0.68

 exorc

0.67

Ig

0.66

Fou

0.65

 contrad

0.65

 aest

0.64

 verses

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.