INDEX

Explanations

words related to political figures and locations

oai_token-act-pair · gpt-3.5-turbo

references to political affiliations and local demographics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

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

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 domestically

-0.74

 Bihar

-0.70

 Australians

-0.69

 Vienna

-0.68

 Countries

-0.68

Sri

-0.68

 Scotland

-0.68

 Australian

-0.68

 Kiev

-0.68

 Gujarat

-0.68

POSITIVE LOGITS

thouse

0.70

itol

0.69

ovember

0.69

uncture

0.66

payers

0.62

icent

0.62

ribune

0.62

 specializing

0.61

ļéĨĴ

0.61

unct

0.60

Activations Density 0.615%

words related to political figures and locations

references to political affiliations and local demographics

No Comments

No Known Activations