INDEX

Explanations

metadata or editorial information within the text, such as editing instructions or revision history

oai_token-act-pair · gpt-3.5-turbo

references to sources or citations in the text

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

wagen

-0.80

merce

-0.72

emouth

-0.72

rontal

-0.70

edIn

-0.70

mable

-0.69

arling

-0.69

eday

-0.68

jriwal

-0.67

oke

-0.67

POSITIVE LOGITS

 edit

1.37

?]

1.32

ËĪ

1.07

!]

1.02

Pg

0.99

:]

0.97

via

0.90

0.89

actionDate

0.89

...]

0.88

Activations Density 0.025%

metadata or editorial information within the text, such as editing instructions or revision history

references to sources or citations in the text

No Known Activations