INDEX

Explanations

mentions of locations

oai_token-act-pair · gpt-3.5-turbo

references to "New," particularly in relation to geographical locations or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

pless

-0.81

ONSORED

-0.80

ï¸ı

-0.74

 sqor

-0.73

ascript

-0.71

uca

-0.70

icent

-0.67

 actionGroup

-0.65

iage

-0.64

aminer

-0.64

POSITIVE LOGITS

 York

1.52

 Zealand

1.49

 Orleans

1.41

 Yorker

1.30

 Yorkers

1.28

 Hampshire

1.19

York

1.18

 Jersey

1.18

 Testament

1.16

 Year

1.10

Activations Density 0.053%

mentions of locations

references to "New," particularly in relation to geographical locations or entities

No Comments

No Known Activations