INDEX

Explanations

terms related to physical damage or impact

oai_token-act-pair · gpt-3.5-turbo

terms related to dental and criminal professions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Object

-0.73

iris

-0.70

 Rated

-0.68

 Brothers

-0.66

ynchronous

-0.66

Flow

-0.66

owl

-0.64

 Sasha

-0.63

ovie

-0.63

Bel

-0.62

POSITIVE LOGITS

 dent

4.34

 Dent

2.51

 dentist

2.13

 dental

2.11

 teeth

1.34

 tooth

1.25

 dermat

1.23

 chirop

1.22

 crim

1.16

 oste

1.16

Activations Density 0.019%

terms related to physical damage or impact

terms related to dental and criminal professions

No Comments

No Known Activations