INDEX

Explanations

the word "surprisingly"

oai_token-act-pair · gpt-3.5-turbo

the phrase "not surprisingly" and its variations

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

abase

-0.85

aily

-0.78

hab

-0.70

uel

-0.70

bor

-0.68

 lineback

-0.66

oreal

-0.65

afety

-0.64

orem

-0.63

enei

-0.63

POSITIVE LOGITS

 srfAttach

0.68

amazon

0.68

 horr

0.67

imaru

0.66

 assuming

0.65

 coincides

0.65

 conclude

0.64

 prompted

0.63

 assume

0.62

ItemThumbnailImage

0.60

Activations Density 0.025%

the word "surprisingly"

the phrase "not surprisingly" and its variations

No Comments

No Known Activations