INDEX

Explanations

phrases related to software or game updates

oai_token-act-pair · gpt-3.5-turbo

references to software patches and updates

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

 Flavoring

-0.67

emouth

-0.61

ccording

-0.61

uron

-0.61

 Galile

-0.60

 VIDEOS

-0.59

 Rivers

-0.58

 Absent

-0.57

 obser

-0.57

 âĸº

-0.56

POSITIVE LOGITS

work

1.27

notes

1.01

oul

0.95

ions

0.88

ing

0.85

ioned

0.84

ogue

0.82

ting

0.81

intosh

0.79

cake

0.79

Activations Density 0.052%

phrases related to software or game updates

references to software patches and updates

No Comments

No Known Activations