INDEX

Explanations

names of individuals

oai_token-act-pair · gpt-3.5-turbo

names of individuals or proper nouns

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 polar

-0.63

 mathemat

-0.62

ccording

-0.62

soDeliveryDate

-0.61

 Citiz

-0.60

cyt

-0.59

 polarization

-0.59

 modifiers

-0.59

 strand

-0.58

 polarized

-0.58

POSITIVE LOGITS

Jr

1.30

Sr

1.01

III

0.92

chuk

0.85

stadt

0.82

wine

0.81

velt

0.80

holm

0.79

QC

0.78

ault

0.77

Activations Density 0.390%

names of individuals

names of individuals or proper nouns

No Comments

No Known Activations