INDEX

Explanations

references to legal or criminal activities

oai_token-act-pair · gpt-3.5-turbo

action and events related to crime and conflict

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

Flickr

-0.75

 Flickr

-0.70

 keynote

-0.69

 Himself

-0.68

 deserves

-0.67

oubtedly

-0.66

 himself

-0.64

uably

-0.62

mittedly

-0.61

omics

-0.58

POSITIVE LOGITS

 themselves

1.27

selves

1.17

 respectively

1.02

 respective

0.95

 their

0.92

 selves

0.90

 helmets

0.82

 careers

0.80

 apiece

0.80

 jointly

0.78

Activations Density 0.738%

references to legal or criminal activities

action and events related to crime and conflict

No Comments

No Known Activations