INDEX

Explanations

references to ongoing television programs

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

mentions of ongoing or continuing events

oai_token-act-pair · gpt-3.5-turbo

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

 contacted

-0.64

liest

-0.62

 cruc

-0.60

unker

-0.60

 WARN

-0.59

 glac

-0.59

 largeDownload

-0.58

 ballistic

-0.58

 tested

-0.58

 preceded

-0.57

POSITIVE LOGITS

 Flavoring

0.79

igue

0.72

ertodd

0.72

 Allaah

0.70

uristic

0.69

fusc

0.65

tyard

0.65

artment

0.65

lace

0.64

wcs

0.64

Activations Density 0.000%

references to ongoing television programs

mentions of ongoing or continuing events

No Comments

No Known Activations