INDEX

Explanations

quotations from news articles or statements

oai_token-act-pair · gpt-3.5-turbo

content that involves quotes from various sources or individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

iencies

-0.69

olit

-0.68

together

-0.64

 Subtle

-0.64

hetics

-0.61

imental

-0.60

meric

-0.60

 Created

-0.59

 âĸº

-0.59

 Lots

-0.58

POSITIVE LOGITS

 saying

1.87

 telling

1.49

 stating

1.43

 Saying

1.35

 predicting

1.27

 remark

1.26

 insisting

1.26

 suggesting

1.24

 asserting

1.22

 claiming

1.22

Activations Density 0.073%

quotations from news articles or statements

content that involves quotes from various sources or individuals

No Comments

No Known Activations