INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Koen

-0.81

 foss

-0.74

 basil

-0.69

Ort

-0.64

 targ

-0.62

 nost

-0.60

pport

-0.59

address

-0.59

caf

-0.59

ulton

-0.58

POSITIVE LOGITS

 "$:/

0.78

hers

0.71

rib

0.69

 Bleach

0.66

 Mayhem

0.66

berman

0.65

0.63

 Younger

0.63

imbabwe

0.62

rim

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.