INDEX

Explanations

terms related to political and economic discussions, including actions such as robbing, enriching, screwing up, legislation, sold to, and support

oai_token-act-pair · gpt-3.5-turbo

Triggered by @bot

references to financial exploitation and the effects of government policies on the economy

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

ayson

-0.81

 observ

-0.74

thodox

-0.69

 Ange

-0.67

Sensor

-0.66

 consistency

-0.66

 sexes

-0.64

Ern

-0.63

Men

-0.61

 Perspect

-0.61

POSITIVE LOGITS

 bailout

1.37

 trillions

1.25

 giveaways

1.25

 privatization

1.24

 taxpayer

1.20

 subsidies

1.18

 giveaway

1.14

 wasteful

1.14

 bail

1.11

 privat

1.08

Activations Density 0.464%

terms related to political and economic discussions, including actions such as robbing, enriching, screwing up, legislation, sold to, and support

references to financial exploitation and the effects of government policies on the economy

No Known Activations