INDEX

Explanations

personal pronouns and verbs indicating personal actions

oai_token-act-pair · gpt-3.5-turbo

references to personal experiences and relationships

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

atory

-0.61

 Millennium

-0.60

Gad

-0.60

 Innocent

-0.59

 Wald

-0.58

 endorsement

-0.55

 Globe

-0.55

 Aston

-0.55

 Emerson

-0.54

 Amen

-0.53

POSITIVE LOGITS

'll

1.29

've

1.25

're

1.22

'd

1.10

 dunno

1.05

'm

0.97

 haven

0.96

can

0.96

 forgot

0.93

don

0.92

Activations Density 0.544%

personal pronouns and verbs indicating personal actions

references to personal experiences and relationships

No Comments

No Known Activations