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

pmwiki

-0.85

ledge

-0.83

ibaba

-0.74

staking

-0.73

ioned

-0.71

secut

-0.70

kefeller

-0.68

stairs

-0.68

alties

-0.65

places

-0.65

POSITIVE LOGITS

 Hurricanes

0.68

 Euph

0.68

 Palin

0.67

Eh

0.65

 Cummings

0.64

urai

0.64

AH

0.63

Thu

0.62

CIS

0.62

Osh

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.