INDEX

Explanations

the word "Sans"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "Sans" and related terms

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

è¦ļéĨĴ

-0.68

UST

-0.67

ETA

-0.61

arious

-0.61

Toy

-0.61

 Weinstein

-0.61

razil

-0.61

ership

-0.61

æ³

-0.61

ocker

-0.60

POSITIVE LOGITS

 Sans

1.28

 sans

1.10

 fonts

0.85

font

0.84

cha

0.80

chal

0.78

faces

0.77

seless

0.76

creen

0.76

crit

0.75

Activations Density 0.010%

the word "Sans"

occurrences of the word "Sans" and related terms

No Comments

No Known Activations