INDEX

Explanations

phrases indicating negation or absence

oai_token-act-pair · gpt-3.5-turbo

phrases that repeatedly reference the word "in" along with contexts indicating location or inclusion

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

DragonMagazine

-0.79

soDeliveryDate

-0.70

rather

-0.68

hai

-0.66

Fig

-0.66

igi

-0.65

adder

-0.63

figured

-0.63

nee

-0.63

GV

-0.60

POSITIVE LOGITS

 anymore

1.68

nor

1.38

any

1.19

 anything

1.12

 anybody

1.12

 slightest

1.08

 anywhere

1.05

 whatsoever

1.04

 anyone

1.03

 necessarily

0.99

Activations Density 0.310%

phrases indicating negation or absence

phrases that repeatedly reference the word "in" along with contexts indicating location or inclusion

No Comments

No Known Activations