INDEX

Explanations

phrases related to crimes or committing illegal activities

oai_token-act-pair · gpt-3.5-turbo

the occurrences of the word "crimes."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 crest

-0.65

 scholarship

-0.65

 brood

-0.62

 swast

-0.62

 bailout

-0.61

 burner

-0.60

 registry

-0.60

 delegation

-0.59

 Scholarship

-0.59

 Buddy

-0.59

POSITIVE LOGITS

imes

5.10

ime

2.63

IME

2.06

imer

1.30

etimes

1.19

oday

1.16

imates

1.13

unes

1.13

ivities

1.10

ima

1.07

Activations Density 0.016%

phrases related to crimes or committing illegal activities

the occurrences of the word "crimes."

No Comments

No Known Activations