INDEX

Explanations

terms related to organized crime activities

oai_token-act-pair · gpt-3.5-turbo

references to organized crime or mob-related themes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

éĸ

-0.84

ANCE

-0.74

tenance

-0.71

______

-0.70

ces

-0.69

gain

-0.68

/+

-0.64

irst

-0.63

 Situation

-0.63

LEASE

-0.63

POSITIVE LOGITS

mob

1.11

 mobs

1.06

Mob

0.88

sters

0.82

busters

0.80

mob

0.79

olean

0.78

iles

0.75

ographed

0.75

 thugs

0.74

Activations Density 0.005%

terms related to organized crime activities

references to organized crime or mob-related themes

No Comments

No Known Activations