INDEX

Explanations

the word "Meghan" with variations in spelling and context

oai_token-act-pair · gpt-3.5-turbo

references to Afghanistan

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

 slack

-0.73

EV

-0.66

 Hubble

-0.65

 Kafka

-0.62

 predictive

-0.61

Reviewed

-0.61

 probable

-0.61

ection

-0.61

 paid

-0.60

 custod

-0.59

POSITIVE LOGITS

ghan

1.46

istan

1.37

igan

1.05

istani

1.03

igans

1.00

awan

0.91

thur

0.90

isma

0.89

wen

0.89

abal

0.88

Activations Density 0.004%

the word "Meghan" with variations in spelling and context

references to Afghanistan

No Comments

No Known Activations