INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

venge

-0.82

adelphia

-0.71

ilogy

-0.69

 overwrite

-0.63

omatic

-0.62

 adjustment

-0.62

 atmosp

-0.62

ieth

-0.62

cs

-0.62

 comma

-0.61

POSITIVE LOGITS

patrick

0.74

Amazon

0.69

rad

0.68

armac

0.67

bett

0.66

Electric

0.66

Introduced

0.66

 Parade

0.66

zik

0.66

 McCabe

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.