INDEX

Explanations

the name "Fran" occurring in various contexts

oai_token-act-pair · gpt-3.5-turbo

proper nouns, specifically names of individuals

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

gow

-0.76

eenth

-0.73

ammed

-0.70

 tomat

-0.70

een

-0.68

wrapper

-0.68

essee

-0.67

skinned

-0.65

pter

-0.65

scape

-0.65

POSITIVE LOGITS

 Fran

1.25

 Dres

1.04

Fran

0.98

furt

0.96

zen

0.88

kel

0.87

ãĥ£

0.83

Vog

0.82

opol

0.79

isine

0.78

Activations Density 0.010%

the name "Fran" occurring in various contexts

proper nouns, specifically names of individuals

No Comments

No Known Activations