INDEX

Explanations

phrases related to authority or positions of power

oai_token-act-pair · gpt-3.5-turbo

references to authority figures or concepts of authority

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

Lens

-0.76

kies

-0.76

auder

-0.74

lla

-0.70

irling

-0.69

esta

-0.67

akeru

-0.66

pocket

-0.65

 Taste

-0.65

bart

-0.65

POSITIVE LOGITS

 vested

1.11

 delegated

1.05

 authority

1.04

 conferred

0.89

 granted

0.86

 exercised

0.86

 overseeing

0.84

Reviewer

0.84

 figures

0.82

 nomine

0.80

Activations Density 0.031%

phrases related to authority or positions of power

references to authority figures or concepts of authority

No Comments

No Known Activations