INDEX

Explanations

numerical values followed by "MW" (megawatts)

oai_token-act-pair · gpt-3.5-turbo

references to megawatt measurements and related energy production terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

++++++++++++++++

-0.65

 Castro

-0.64

 Caesar

-0.63

Vit

-0.63

Bone

-0.63

 scratched

-0.62

 Flesh

-0.61

psy

-0.61

voy

-0.61

bart

-0.61

POSITIVE LOGITS

MW

1.38

MW

1.14

LW

1.01

NT

1.01

IED

0.97

LLOW

0.94

NING

0.90

orthy

0.90

OOD

0.90

urst

0.89

Activations Density 0.006%

numerical values followed by "MW" (megawatts)

references to megawatt measurements and related energy production terms

No Comments

No Known Activations