INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

neys

-0.78

espie

-0.73

mates

-0.70

lain

-0.68

 governors

-0.68

Reviewer

-0.68

neath

-0.65

engers

-0.63

aston

-0.62

 Governors

-0.61

POSITIVE LOGITS

insk

0.85

anguages

0.74

 Manufacturer

0.72

icz

0.72

 Drinking

0.67

HRC

0.65

 Haku

0.65

 Suzuki

0.63

oteric

0.63

 acupuncture

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.