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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ģĸ

-0.94

selves

-0.86

\\\\\\\\

-0.73

Klu

-0.70

 disclaimer

-0.68

©¶æ¥µ

-0.66

 Handling

-0.64

 Settlement

-0.64

gery

-0.63

geries

-0.62

POSITIVE LOGITS

osen

0.85

imes

0.77

alus

0.74

anqu

0.73

izoph

0.70

ilant

0.69

å§«

0.67

resent

0.66

audi

0.66

oldown

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.