INDEX

Explanations

phrases related to images or photographs from the Getty Images collection

oai_token-act-pair · gpt-3.5-turbo

occurrences of the brand name 'Getty Images'

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 sediment

-0.73

 misunderstood

-0.67

 particip

-0.65

 timely

-0.65

heartedly

-0.65

 infiltr

-0.64

stood

-0.64

 forgiven

-0.62

 stricken

-0.62

 reinvest

-0.62

POSITIVE LOGITS

sburg

1.74

 Images

1.63

Images

1.18

sburgh

1.07

images

1.02

photo

0.83

Film

0.81

 IMAGES

0.80

Photos

0.79

 Features

0.77

Activations Density 0.024%

phrases related to images or photographs from the Getty Images collection

occurrences of the brand name 'Getty Images'

No Comments

No Known Activations