INDEX

Explanations

news articles prompting readers to 'Read more'

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "Read" followed by more information or links to articles

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

xon

-0.76

odied

-0.74

IDA

-0.72

ounty

-0.69

TEXTURE

-0.68

IDS

-0.67

EStreamFrame

-0.65

opard

-0.65

channelAvailability

-0.64

-"

-0.63

POSITIVE LOGITS

 aloud

0.91

ahead

0.83

iness

0.81

ied

0.81

 Read

0.81

Write

0.81

sburg

0.80

ying

0.80

just

0.79

ies

0.72

Activations Density 0.011%

news articles prompting readers to 'Read more'

occurrences of the word "Read" followed by more information or links to articles

No Comments

No Known Activations