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

ľ

-0.68

paren

-0.68

ermott

-0.68

®

-0.67

 Chaff

-0.66

 Fiat

-0.65

grave

-0.64

 Frankie

-0.62

ļéĨĴ

-0.62

Flo

-0.61

POSITIVE LOGITS

 closest

0.89

APS

0.64

roxy

0.63

omsky

0.62

athering

0.61

awar

0.61

phone

0.61

Ay

0.61

agy

0.59

eware

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.