INDEX

Explanations

words related to criminal and investigative activities, such as robberies, searches, and bombings

oai_token-act-pair · gpt-3.5-turbo

terms related to crime and incidents

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

asus

-0.66

Stud

-0.61

coat

-0.59

 coat

-0.59

shield

-0.58

 Malk

-0.55

REDACTED

-0.55

ISH

-0.54

 Coat

-0.54

 Canal

-0.54

POSITIVE LOGITS

uggest

1.36

hip

1.21

mith

1.21

hips

1.20

ettings

1.20

poons

1.18

hooting

1.18

uits

1.16

paces

1.14

peed

1.10

Activations Density 0.352%

words related to criminal and investigative activities, such as robberies, searches, and bombings

terms related to crime and incidents

No Comments

No Known Activations