INDEX

Explanations

instances of quantitative values being presented with a range

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a range or scale pertaining to various topics or measurements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

faced

-0.80

mask

-0.79

seek

-0.76

quote

-0.75

isode

-0.73

rongh

-0.72

ratulations

-0.70

lied

-0.69

tailed

-0.68

blance

-0.68

POSITIVE LOGITS

 afar

1.43

 whence

1.17

 thence

1.12

 abroad

0.96

 scratch

0.90

 anywhere

0.90

 somewhere

0.89

 wherever

0.85

 elsewhere

0.84

 inside

0.83

Activations Density 0.291%

instances of quantitative values being presented with a range

phrases indicating a range or scale pertaining to various topics or measurements

No Comments

No Known Activations