INDEX

Explanations

phrases indicating a search for something specific or a desired outcome

oai_token-act-pair · gpt-3.5-turbo

phrases indicating ongoing searches or inquiries

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Democr

-0.76

soever

-0.68

owned

-0.66

ANK

-0.65

minus

-0.64

 racked

-0.63

NPR

-0.62

FactoryReloaded

-0.62

 notwithstanding

-0.61

own

-0.61

POSITIVE LOGITS

 ways

1.11

 clues

1.10

 answers

1.05

 solutions

1.02

 scapego

0.96

 excuses

0.92

 alternatives

0.92

 explanations

0.91

 opportunities

0.89

 inspiration

0.87

Activations Density 0.076%

phrases indicating a search for something specific or a desired outcome

phrases indicating ongoing searches or inquiries

No Comments

No Known Activations