INDEX

Explanations

words related to news, updates, and services

oai_token-act-pair · gpt-3.5-turbo

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

minster

-0.71

etheless

-0.70

venth

-0.67

iom

-0.61

cause

-0.60

zon

-0.57

lihood

-0.57

veland

-0.57

oken

-0.57

poral

-0.55

POSITIVE LOGITS

 please

1.20

 visit

1.09

please

1.07

 click

1.02

 CLICK

0.98

see

0.92

 check

0.87

 consult

0.87

 subscribe

0.84

 PLEASE

0.80

Activations Density 0.068%

words related to news, updates, and services

No Comments

No Known Activations