INDEX

Explanations

instances of the word "privileges."

oai_token-act-pair · gpt-3.5-turbo

references to privileges and legal permissions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

source

-0.83

tra

-0.77

TE

-0.73

sil

-0.73

sell

-0.73

agh

-0.73

 Archangel

-0.73

seller

-0.72

leaf

-0.72

Ø§ÙĦ

-0.72

POSITIVE LOGITS

 privileges

1.63

 entitle

1.19

 permissions

1.14

ilege

1.09

 privilege

1.05

ileged

0.92

 conferred

0.91

 credentials

0.87

holders

0.87

 exemptions

0.87

Activations Density 0.008%

instances of the word "privileges."

references to privileges and legal permissions

No Comments

No Known Activations