INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

references to places or regions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Py

-0.66

 mutual

-0.65

 glob

-0.64

 burg

-0.62

 origin

-0.62

 revolving

-0.61

 Pixie

-0.60

ly

-0.60

 predatory

-0.60

 covering

-0.59

POSITIVE LOGITS

ada

4.52

adas

3.08

ADA

2.29

ado

1.68

adian

1.64

adan

1.43

adi

1.34

uda

1.33

ador

1.29

ade

1.28

Activations Density 0.020%

proper nouns

references to places or regions

No Comments

No Known Activations