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

levard

-0.74

airs

-0.69

roma

-0.69

uda

-0.69

 Jackets

-0.66

ikes

-0.64

illary

-0.62

Zone

-0.62

IJ

-0.61

ienne

-0.61

POSITIVE LOGITS

 cutoff

0.67

GGGG

0.63

 Scrolls

0.62

GGGGGGGG

0.61

actionDate

0.61

 Peterson

0.61

#$

0.60

swick

0.59

epad

0.59

BuyableInstoreAndOnline

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.