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

OPLE

-0.84

dayName

-0.83

uve

-0.81

phe

-0.79

IED

-0.79

beit

-0.76

dy

-0.71

IAL

-0.70

kered

-0.70

EEE

-0.69

POSITIVE LOGITS

 Accord

0.76

irez

0.66

Cash

0.61

 Civilization

0.61

Var

0.61

 Rhino

0.61

bye

0.61

balance

0.60

 withdrawals

0.60

Balance

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.