INDEX

Explanations

terms related to law enforcement and investigations

oai_token-act-pair · gpt-3.5-turbo

instances of numbers or percentages in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 honor

-0.67

 Marilyn

-0.64

 honors

-0.60

 Marian

-0.59

 Oprah

-0.57

 honoring

-0.56

.)

-0.55

 AIDS

-0.55

.]

-0.54

 Hispanic

-0.54

POSITIVE LOGITS

etheless

0.89

DragonMagazine

0.87

eworks

0.74

ijn

0.73

ework

0.70

ãĥ¯ãĥ³

0.65

staking

0.63

ÃįÃį

0.63

escription

0.62

 skelet

0.62

Activations Density 1.546%

terms related to law enforcement and investigations

instances of numbers or percentages in the text

No Comments

No Known Activations