INDEX

Explanations

numbers and financial terms

oai_token-act-pair · gpt-3.5-turbo

affirmations and significant numerical values

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ectomy

-0.67

keye

-0.57

clave

-0.57

 Norman

-0.56

oise

-0.53

silver

-0.53

weights

-0.52

 latitude

-0.52

uro

-0.52

osi

-0.52

POSITIVE LOGITS

anwhile

0.86

[/

0.71

 coron

0.70

xon

0.69

xual

0.69

 dunno

0.68

dow

0.68

etheless

0.66

jri

0.64

ESE

0.62

Activations Density 0.156%

numbers and financial terms

affirmations and significant numerical values

No Comments

No Known Activations