INDEX

Explanations

references to political figures

oai_token-act-pair · gpt-3.5-turbo

conversational elements and expressions of consideration or preference

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

.)

-0.69

AFP

-0.66

 damning

-0.64

 Shutterstock

-0.64

 awaits

-0.63

MSM

-0.61

 Poverty

-0.61

 powerless

-0.61

 ACLU

-0.60

 centuries

-0.60

POSITIVE LOGITS

laughs

1.08

Laughs

1.00

initely

0.94

['

0.92

entimes

0.88

yss

0.87

0.83

laughter

0.81

ovie

0.79

icularly

0.78

Activations Density 0.659%

references to political figures

conversational elements and expressions of consideration or preference

No Comments

No Known Activations