INDEX

Explanations

names of individuals

oai_token-act-pair · gpt-3.5-turbo

references to personal relationships and social interactions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 antitrust

-0.70

£ı

-0.70

 mathemat

-0.64

 indict

-0.63

 optics

-0.63

 misleading

-0.63

 circumvent

-0.62

 billionaires

-0.60

 surfaces

-0.60

Russia

-0.60

POSITIVE LOGITS

 roomm

0.98

 babys

0.95

 girls

0.90

 roommate

0.89

 girl

0.89

 classmates

0.88

 classmate

0.85

 youngest

0.83

unts

0.83

 sisters

0.83

Activations Density 1.078%

names of individuals

references to personal relationships and social interactions

No Comments

No Known Activations