INDEX

Explanations

multiples or various forms of "fold" increases or occurrences

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the concept of "folding" or "layers."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

ampires

-0.73

EA

-0.73

culus

-0.72

ahoo

-0.71

IRO

-0.71

angan

-0.70

assetsadobe

-0.70

APS

-0.70

FTWARE

-0.69

ody

-0.68

POSITIVE LOGITS

fold

1.17

 fold

0.86

 increase

0.83

theless

0.73

edly

0.73

shr

0.72

forward

0.72

step

0.71

 reliance

0.67

 enlarg

0.66

Activations Density 0.024%

multiples or various forms of "fold" increases or occurrences

phrases indicating the concept of "folding" or "layers."

No Comments

No Known Activations