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

 guiActiveUn

-0.92

meier

-0.79

rak

-0.73

ulz

-0.70

chall

-0.69

Sov

-0.68

 footing

-0.68

 offending

-0.67

arius

-0.67

sing

-0.67

POSITIVE LOGITS

 Prohibition

0.72

GREEN

0.69

Vie

0.67

 Appalach

0.65

 Habit

0.64

WASHINGTON

0.62

 IMAGES

0.61

————

0.61

ORTS

0.61

Wem

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.