INDEX

Explanations

the word "House" followed by another word or phrase

oai_token-act-pair · gpt-3.5-turbo

references to specific titles of books and shows

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

istically

-0.80

estamp

-0.76

istical

-0.71

istic

-0.70

asm

-0.69

onal

-0.68

inez

-0.63

ï¸ı

-0.62

oth

-0.62

oshop

-0.61

POSITIVE LOGITS

wives

1.53

keeping

1.41

wife

1.40

hold

1.14

holders

1.04

maid

1.03

warming

1.02

holder

0.98

keepers

0.97

plant

0.94

Activations Density 0.036%

the word "House" followed by another word or phrase

references to specific titles of books and shows

No Comments

No Known Activations