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

 Empress

-0.82

 Norn

-0.72

 Puppet

-0.67

 Knot

-0.66

İĭ

-0.66

hess

-0.63

 Pengu

-0.63

 capitals

-0.61

 Dagger

-0.61

 Accord

-0.61

POSITIVE LOGITS

drug

0.70

catentry

0.69

qus

0.67

baugh

0.65

Abu

0.64

uro

0.63

endment

0.61

iday

0.61

pard

0.60

achment

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.