INDEX

Explanations

dates in the format "Month day" with high activation values

oai_token-act-pair · gpt-3.5-turbo

dates, specifically in March

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

lder

-0.68

kB

-0.67

 tray

-0.66

itaire

-0.66

tox

-0.63

gone

-0.62

 constitu

-0.61

 quot

-0.59

 unnecess

-0.59

folios

-0.59

POSITIVE LOGITS

 Madness

1.20

nard

0.91

riage

0.91

0.88

rd

0.84

ing

0.84

0.82

onna

0.81

 Hare

0.80

yard

0.79

Activations Density 0.023%

dates in the format "Month day" with high activation values

dates, specifically in March

No Comments

No Known Activations