INDEX

Explanations

words related to decorative items or awards

oai_token-act-pair · gpt-3.5-turbo

references to ribbons and stripes in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

regor

-0.86

CVE

-0.80

efer

-0.75

arios

-0.73

othy

-0.71

aeper

-0.70

ormon

-0.69

 Aval

-0.69

otrop

-0.68

 burgl

-0.68

POSITIVE LOGITS

 ribbon

1.38

 stripes

1.27

 stripe

1.19

 Ribbon

1.19

 strand

0.86

 striped

0.82

 feathers

0.76

backer

0.76

accompan

0.75

 strands

0.75

Activations Density 0.024%

words related to decorative items or awards

references to ribbons and stripes in various contexts

No Comments

No Known Activations