INDEX

Explanations

pairings or relationships between different individuals

oai_token-act-pair · gpt-3.5-turbo

instances of the word "and" to indicate connections or relationships

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

artifacts

-0.83

pmwiki

-0.79

system

-0.74

attribute

-0.72

vernment

-0.67

 captcha

-0.67

arer

-0.67

Examples

-0.67

oting

-0.65

omsky

-0.65

POSITIVE LOGITS

 Ellie

1.19

 Michelle

1.18

 Daisy

1.17

 Fiona

1.17

 Donna

1.17

 Denise

1.16

 Tammy

1.14

 Jackie

1.13

 Veronica

1.12

 Morty

1.12

Activations Density 0.124%

pairings or relationships between different individuals

instances of the word "and" to indicate connections or relationships

No Comments

No Known Activations