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

 charms

-0.79

Ari

-0.74

 wishes

-0.66

 Bere

-0.65

SIG

-0.65

 Grande

-0.64

 Gard

-0.64

 enqu

-0.64

 Sammy

-0.63

Pis

-0.62

POSITIVE LOGITS

fter

0.90

eele

0.83

host

0.80

 ingred

0.79

usha

0.75

pmwiki

0.75

azy

0.74

ebted

0.74

Guest

0.74

DAQ

0.73

Activations Density 0.000%

No Known Activations

This feature has no known activations.