INDEX

Explanations

references to locations or geographical places, particularly those with 'New' in the name

oai_token-act-pair · gpt-3.5-turbo

mentions of the state "New" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ï¸ı

-0.86

ONSORED

-0.85

Reloaded

-0.78

 sqor

-0.78

 actionGroup

-0.78

ILA

-0.77

ascript

-0.74

xual

-0.74

 DRAG

-0.72

ickr

-0.69

POSITIVE LOGITS

 York

1.39

 Zealand

1.33

 Hampshire

1.24

 Orleans

1.22

 Jersey

1.11

bie

1.04

bies

1.03

 Yorker

1.02

 Yorkers

0.99

egg

0.98

Activations Density 0.052%

references to locations or geographical places, particularly those with 'New' in the name

mentions of the state "New" in various contexts

No Comments

No Known Activations