INDEX

Explanations

medical and health-related treatments

oai_token-act-pair · gpt-3.5-turbo

references to medical treatments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

---------------

-0.72

de

-0.72

ord

-0.72

externalActionCode

-0.70

EG

-0.66

DEP

-0.64

 Huff

-0.64

zo

-0.64

earth

-0.63

owned

-0.61

POSITIVE LOGITS

 treatments

1.27

 therapies

1.03

apons

0.92

terday

0.89

adeon

0.87

ayers

0.87

olutions

0.86

poons

0.85

kinson

0.85

ktop

0.85

Activations Density 0.012%

medical and health-related treatments

references to medical treatments

No Comments

No Known Activations