INDEX

Explanations

articles containing a video and/or emphasizing the act of watching a video

oai_token-act-pair · gpt-3.5-turbo

references to videos and live content

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

insula

-0.67

RIC

-0.66

nir

-0.66

²¾

-0.66

Syl

-0.66

Ħ¢

-0.65

isma

-0.64

XL

-0.64

 reusable

-0.64

 âĵĺ

-0.63

POSITIVE LOGITS

 unfold

1.14

 VIDEOS

1.00

ideos

0.91

 replay

0.86

 videos

0.85

 tapes

0.85

 closely

0.81

 movies

0.81

 broadcasts

0.80

 films

0.80

Activations Density 0.157%

articles containing a video and/or emphasizing the act of watching a video

references to videos and live content

No Comments

No Known Activations