INDEX

Explanations

phrases indicating location, especially places where people stay temporarily

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate locations, particularly involving the word "at."

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

ratulations

-0.79

Pac

-0.75

archy

-0.74

Î³

-0.73

alpha

-0.72

aceutical

-0.70

iOS

-0.69

HTML

-0.69

zyk

-0.69

INAL

-0.68

POSITIVE LOGITS

 least

1.16

onement

1.04

 hotels

0.98

 Disneyland

0.91

 cafes

0.90

 airports

0.90

 home

0.90

 restaurants

0.88

 Costco

0.87

 CafÃ©

0.81

Activations Density 0.232%

phrases indicating location, especially places where people stay temporarily

phrases that indicate locations, particularly involving the word "at."

No Comments

No Known Activations