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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 heights

-0.77

 overtime

-0.67

 tradem

-0.63

 exha

-0.61

 Jaguars

-0.60

 earthqu

-0.60

 helicop

-0.60

Â¯

-0.59

 sche

-0.59

Þ

-0.59

POSITIVE LOGITS

draw

0.76

metics

0.74

atile

0.72

oke

0.72

cade

0.72

naire

0.70

oking

0.69

okes

0.68

rame

0.68

ivation

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.