INDEX

Explanations

locations and dates mentioned in news articles

oai_token-act-pair · gpt-3.5-turbo

punctuations and commas in dates and locations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

CONCLUS

-0.68

ãĥĥãĥī

-0.67

 impulse

-0.58

estone

-0.57

worldly

-0.57

 Loft

-0.56

ophe

-0.55

©¶æ¥µ

-0.55

,—

-0.54

ãĤ£

-0.54

POSITIVE LOGITS

 circa

0.95

 pictured

0.79

 REUTERS

0.73

window

0.70

 Sept

0.69

Feb

0.69

 Calif

0.68

 photographed

0.68

 Photo

0.67

 Photographer

0.65

Activations Density 0.078%

locations and dates mentioned in news articles

punctuations and commas in dates and locations

No Comments

No Known Activations

locations and dates mentioned in news articles

punctuations and commas in dates and locations

No Comments

No Known Activations