INDEX

Explanations

phrases related to comparison or contrast

oai_token-act-pair · gpt-3.5-turbo

conjunctions and phrases indicating relationships or connections between ideas

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

coon

-0.70

IONS

-0.66

Fuck

-0.65

hire

-0.64

ventory

-0.63

pired

-0.62

umbn

-0.62

åŃĲ

-0.61

owered

-0.61

°

-0.61

POSITIVE LOGITS

 occurs

1.30

 tends

1.30

 varies

1.29

 arises

1.27

 involves

1.27

 differs

1.24

 manifests

1.22

 persists

1.22

 contributes

1.17

 deserves

1.16

Activations Density 0.443%

phrases related to comparison or contrast

conjunctions and phrases indicating relationships or connections between ideas

No Comments

No Known Activations