INDEX

Explanations

references to different perspectives or viewpoints

oai_token-act-pair · gpt-3.5-turbo

phrases indicating perspectives or viewpoints

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Guinness

-0.68

 Buzz

-0.67

 congratulations

-0.67

 Seal

-0.66

 Senators

-0.64

 Candy

-0.64

Buzz

-0.63

 Amen

-0.62

 Congratulations

-0.62

 Hunts

-0.62

POSITIVE LOGITS

 basis

1.18

ysical

1.04

 standpoint

1.00

entric

0.96

ocial

0.94

iological

0.93

 perspect

0.93

 behavi

0.92

 interpersonal

0.91

 pragmatic

0.91

Activations Density 0.648%

references to different perspectives or viewpoints

phrases indicating perspectives or viewpoints

No Comments

No Known Activations