INDEX

Explanations

video titles containing the phrase "MUST WATCH."

oai_token-act-pair · gpt-3.5-turbo

the phrase "MUST WATCH" and related calls to action regarding viewing content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

urally

-0.71

 academia

-0.70

 bats

-0.70

 sucker

-0.67

ura

-0.67

jury

-0.66

 pulp

-0.66

Dee

-0.66

 nowhere

-0.66

utra

-0.64

POSITIVE LOGITS

 WATCH

1.24

 Thumbnails

0.97

 VIDEOS

0.97

 WATCHED

0.93

 Watching

0.88

FILE

0.82

]}

0.80

IFF

0.80

CLASSIFIED

0.79

LIST

0.79

Activations Density 0.006%

video titles containing the phrase "MUST WATCH."

the phrase "MUST WATCH" and related calls to action regarding viewing content

No Comments

No Known Activations