INDEX

Explanations

references to knowledge or information

oai_token-act-pair · gpt-3.5-turbo

instances of the word "know" and its variations, indicating a focus on knowledge or awareness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Mehran

-0.76

Pax

-0.68

 decorations

-0.61

 cartoons

-0.59

 persuasion

-0.58

 voucher

-0.58

 Statue

-0.57

 rever

-0.57

 Singer

-0.57

 frivolous

-0.56

POSITIVE LOGITS

lege

1.53

ledged

1.44

ledge

1.31

LED

1.25

led

1.08

ingly

0.99

ability

0.96

how

0.92

yers

0.88

ably

0.88

Activations Density 0.044%

references to knowledge or information

instances of the word "know" and its variations, indicating a focus on knowledge or awareness

No Comments

No Known Activations