INDEX

Explanations

keywords related to photographs/photos

oai_token-act-pair · gpt-3.5-turbo

references to photos

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

bilt

-0.84

ressive

-0.71

uyomi

-0.70

ensable

-0.69

inence

-0.69

vernment

-0.68

merce

-0.67

 Rutherford

-0.66

 Clause

-0.66

ricular

-0.64

POSITIVE LOGITS

ynthesis

1.50

ynt

1.19

ensitive

1.03

photos

1.00

journal

0.98

 depicting

0.86

Photos

0.86

 Phot

0.86

 photographed

0.84

 snapped

0.84

Activations Density 0.026%

keywords related to photographs/photos

references to photos

No Comments

No Known Activations