INDEX

Explanations

references to governmental or political roles and structures, specifically related to cabinet positions

oai_token-act-pair · gpt-3.5-turbo

references to "cabinet" in a political context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

jas

-0.70

Season

-0.64

mol

-0.64

Inv

-0.64

baugh

-0.62

inv

-0.61

Â¶

-0.60

sson

-0.60

Ur

-0.60

 validate

-0.59

POSITIVE LOGITS

 cabinet

4.03

 Cabinet

3.37

 cabinets

2.86

 ministerial

1.54

 drawer

1.49

abin

1.31

 ministers

1.30

 closet

1.29

 Ministers

1.19

 minister

1.17

Activations Density 0.019%

references to governmental or political roles and structures, specifically related to cabinet positions

references to "cabinet" in a political context

No Comments

No Known Activations