INDEX

Explanations

mentions of images

oai_token-act-pair · gpt-3.5-turbo

references to images or visual media

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 recess

-0.77

Interstitial

-0.74

 usher

-0.69

 cones

-0.67

 legislature

-0.66

 multiplication

-0.65

 spokes

-0.63

 educate

-0.62

 holog

-0.61

 enlight

-0.61

POSITIVE LOGITS

addons

0.90

lihood

0.89

SourceFile

0.79

?]

0.74

"]=>

0.73

claim

0.70

']

0.70

rahim

0.69

clip

0.67

aved

0.65

Activations Density 0.000%

mentions of images

references to images or visual media

No Comments

No Known Activations