INDEX

Explanations

instances of objects or entities physically landing somewhere

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "landed."

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

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

inki

-0.83

externalActionCode

-0.83

Cola

-0.73

nature

-0.71

attribute

-0.70

ciplinary

-0.70

Downloadha

-0.69

ope

-0.69

att

-0.67

fore

-0.65

POSITIVE LOGITS

 landed

1.02

 landing

0.89

 lands

0.80

ãĤ¤ãĥĪ

0.76

 land

0.74

 Landing

0.71

 blows

0.67

lehem

0.66

nav

0.66

terson

0.65

Activations Density 0.008%

instances of objects or entities physically landing somewhere

occurrences of the word "landed."

No Comments

No Known Activations