INDEX

Explanations

mentions of the word "privilege."

oai_token-act-pair · gpt-3.5-turbo

references to the concept of privilege

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ood

-0.80

Interstitial

-0.73

ndra

-0.66

GH

-0.65

ailed

-0.62

ail

-0.61

 chaotic

-0.61

agh

-0.60

quarter

-0.60

eros

-0.60

POSITIVE LOGITS

ilege

1.57

 privilege

1.36

ileged

1.17

 privileges

0.95

 Priv

0.77

itism

0.77

 afforded

0.75

 privileged

0.75

 privile

0.72

ilial

0.71

Activations Density 0.007%

mentions of the word "privilege."

references to the concept of privilege

No Comments

No Known Activations