INDEX

Explanations

reports of criminal or law enforcement activities

oai_token-act-pair · gpt-3.5-turbo

references to law enforcement and authorities in the context of criminal activities

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

ictional

-0.66

DragonMagazine

-0.60

PsyNetMessage

-0.58

 torch

-0.58

DonaldTrump

-0.57

noon

-0.56

iatus

-0.55

 tripod

-0.55

osate

-0.53

 alphabet

-0.53

POSITIVE LOGITS

 said

1.20

 says

0.99

said

0.99

 told

0.98

 wrote

0.89

 tells

0.87

 announced

0.86

 Says

0.85

 confirmed

0.84

 explained

0.83

Activations Density 0.250%

reports of criminal or law enforcement activities

references to law enforcement and authorities in the context of criminal activities

No Comments

No Known Activations