INDEX

Explanations

medical or investigative terms and activities

oai_token-act-pair · gpt-3.5-turbo

references to specific subjects or details related to investigations or assessments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

itia

-0.74

NB

-0.68

 Shutterstock

-0.66

ouble

-0.65

vice

-0.64

vest

-0.64

book

-0.64

raud

-0.62

ey

-0.62

âĿ

-0.62

POSITIVE LOGITS

 extent

1.50

 exact

1.42

 precise

1.41

 origin

1.27

 severity

1.27

 whereabouts

1.27

 nature

1.26

 specifics

1.23

 identities

1.23

 origins

1.21

Activations Density 0.253%

medical or investigative terms and activities

references to specific subjects or details related to investigations or assessments

No Comments

No Known Activations