INDEX

Explanations

the pattern "x-user-generated," where x can be any word

oai_token-act-pair · gpt-3.5-turbo

references to user-generated content and its impact on various contexts

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

 Downs

-0.68

 bras

-0.68

 sunset

-0.67

 Burr

-0.66

 Dunk

-0.65

 beads

-0.65

 Canaan

-0.65

 Maiden

-0.65

 Pompe

-0.63

 Bullets

-0.63

POSITIVE LOGITS

friendly

1.48

driven

1.40

generated

1.39

centric

1.30

centered

1.29

oriented

1.26

controlled

1.26

focused

1.25

operated

1.17

facing

1.16

Activations Density 0.059%

the pattern "x-user-generated," where x can be any word

references to user-generated content and its impact on various contexts

No Comments

No Known Activations