INDEX

Explanations

names and titles of individuals

oai_token-act-pair · gpt-3.5-turbo

references to leadership and authority figures

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

Cub

-0.84

PLAY

-0.76

Spanish

-0.69

 Beacon

-0.69

 Sounders

-0.69

prototype

-0.66

 Zurich

-0.65

 Becky

-0.65

 Havana

-0.64

Miami

-0.64

POSITIVE LOGITS

 Sharma

1.52

 Singh

1.52

 Gupta

1.48

Rao

1.43

 Devi

1.41

 Kumar

1.41

 Patel

1.35

Yad

1.28

pta

1.26

 Gandhi

1.26

Activations Density 0.256%

names and titles of individuals

references to leadership and authority figures

No Comments

No Known Activations