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

utenant

-0.80

 Galile

-0.71

Gre

-0.69

dfx

-0.68

 Romans

-0.66

ronics

-0.66

 Turks

-0.66

erence

-0.64

ution

-0.62

 Spani

-0.62

POSITIVE LOGITS

ãĤ´ãĥ³

0.73

estern

0.73

 cert

0.67

 Cert

0.66

pport

0.66

aud

0.65

ascus

0.64

 athlet

0.64

 Beir

0.64

bats

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.