INDEX

Explanations

small numerical values

oai_token-act-pair · gpt-3.5-turbo

quantifiable data and statistics related to group sizes or opinions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ocracy

-0.67

FactoryReloaded

-0.63

 Stamp

-0.60

 Dictionary

-0.60

Chapter

-0.60

Magn

-0.60

uyomi

-0.59

oken

-0.58

cking

-0.58

tnc

-0.57

POSITIVE LOGITS

 survived

0.88

teen

0.88

 withdrew

0.83

 succeeded

0.83

 succumbed

0.82

 were

0.80

 stayed

0.80

 went

0.80

 resulted

0.78

 lasted

0.78

Activations Density 0.307%

small numerical values

quantifiable data and statistics related to group sizes or opinions

No Comments

No Known Activations