INDEX

Explanations

references to specific entities or organizations with a 'C' followed by a number

oai_token-act-pair · gpt-3.5-turbo

references to a particular case or subject denoted by 'C'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 GOODMAN

-0.76

uyomi

-0.75

 Tinder

-0.74

ãĥĹ

-0.71

wagen

-0.70

 Wrestle

-0.70

EStream

-0.69

Magikarp

-0.68

ItemTracker

-0.67

 Yamaha

-0.66

POSITIVE LOGITS

overed

1.28

abbage

1.11

ursor

1.11

rouch

1.11

oding

1.10

aught

1.09

URRENT

1.05

olly

1.05

ottage

1.05

resp

1.05

Activations Density 0.052%

references to specific entities or organizations with a 'C' followed by a number

references to a particular case or subject denoted by 'C'

No Comments

No Known Activations