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

igion

-0.88

ãĤ£

-0.79

ãĥĨãĤ£

-0.74

è¦ļéĨĴ

-0.72

ãĥĥãĥī

-0.72

icter

-0.72

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.71

aterial

-0.68

stals

-0.66

 Sovere

-0.65

POSITIVE LOGITS

 Cancel

0.77

 cance

0.68

 cancellation

0.63

 escort

0.63

processing

0.62

 Form

0.62

][

0.61

 postp

0.61

 renovations

0.60

aming

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.