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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

oka

-0.82

ION

-0.78

oming

-0.70

hern

-0.70

aja

-0.70

ioned

-0.70

oma

-0.69

vez

-0.69

ibles

-0.69

pillar

-0.67

POSITIVE LOGITS

 distance

1.97

 distances

1.41

 Distance

1.26

distance

1.24

Distance

0.99

ridor

0.74

 Situation

0.72

 warmth

0.72

0.70

 mornings

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.