INDEX

Explanations

technical terms and names, possibly related to science and technology

oai_token-act-pair · gpt-3.5-turbo

punctuation marks, particularly commas

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

Submit

-0.72

ances

-0.66

ocy

-0.65

axies

-0.64

joy

-0.63

ocalypse

-0.62

 Statements

-0.61

worldly

-0.60

oday

-0.60

eties

-0.60

POSITIVE LOGITS

 nicknamed

1.41

 which

1.33

 whose

1.23

aka

1.19

 formerly

1.15

 located

1.14

 consisting

1.13

 dubbed

1.12

 pictured

1.11

 named

1.10

Activations Density 0.265%

technical terms and names, possibly related to science and technology

punctuation marks, particularly commas

No Comments

No Known Activations