INDEX

Explanations

phrases indicating alternatives or options

oai_token-act-pair · gpt-3.5-turbo

the word "or" and its connection to various contrasting ideas or options

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ETS

-0.73

Definition

-0.66

vant

-0.64

estro

-0.64

FACE

-0.64

Original

-0.63

Timer

-0.62

"],"

-0.61

Date

-0.61

Ret

-0.61

POSITIVE LOGITS

 alternatively

1.48

 worse

1.26

 else

1.24

 indeed

1.16

 even

1.16

 perhaps

1.15

 maybe

1.14

chard

1.01

 vice

1.01

 downright

0.91

Activations Density 0.088%

phrases indicating alternatives or options

the word "or" and its connection to various contrasting ideas or options

No Comments

No Known Activations