INDEX

Explanations

phrases where someone is listening or being informed about something

oai_token-act-pair · gpt-3.5-turbo

references to auditory perception or listening

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

prime

-0.72

ettel

-0.69

transfer

-0.68

eros

-0.67

EStreamFrame

-0.67

idad

-0.66

wra

-0.65

enture

-0.65

efficient

-0.65

apo

-0.62

POSITIVE LOGITS

 Heard

1.06

 hear

1.05

 Hear

1.00

 heard

0.92

 hearing

0.91

 voices

0.83

eport

0.81

 hears

0.81

 listened

0.79

 Hearing

0.77

Activations Density 0.016%

phrases where someone is listening or being informed about something

references to auditory perception or listening

No Comments

No Known Activations