INDEX

Explanations

phrases related to sharing information or content with others

oai_token-act-pair · gpt-3.5-turbo

instances of sharing information or experiences with others

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

 vacc

-0.72

venant

-0.68

itures

-0.66

orah

-0.65

ucci

-0.64

ischer

-0.62

hemat

-0.61

 excluding

-0.61

 vein

-0.60

zanne

-0.60

POSITIVE LOGITS

 regards

1.14

 regard

1.06

stood

1.04

draw

0.88

 respect

0.87

 coworkers

0.86

 fellow

0.84

 dignity

0.83

 impunity

0.83

us

0.80

Activations Density 0.082%

phrases related to sharing information or content with others

instances of sharing information or experiences with others

No Comments

No Known Activations