INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

zai

-0.92

SEE

-0.82

 Tanz

-0.77

WER

-0.76

Wem

-0.76

oÄŁan

-0.73

 Kush

-0.73

aida

-0.72

ĳå£«

-0.72

wives

-0.72

POSITIVE LOGITS

ategory

0.81

 reservations

0.73

 engagements

0.73

 submissions

0.69

 inhib

0.69

 allegations

0.69

 culminated

0.68

 setbacks

0.67

 torpedo

0.67

 disav

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.