INDEX

Explanations

the name "Loretta" repeated in different contexts and variations

oai_token-act-pair · gpt-3.5-turbo

mentions of specific names, particularly those of public figures in legal 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

ergy

-0.91

shire

-0.89

eric

-0.83

er

-0.78

ships

-0.76

lich

-0.75

bies

-0.74

egg

-0.73

enstein

-0.73

spring

-0.72

POSITIVE LOGITS

arnaev

0.77

 fide

0.74

velength

0.73

ignt

0.70

BILITY

0.69

ĨĴ

0.69

Availability

0.68

 Duty

0.65

utonium

0.64

ioxide

0.64

Activations Density 0.032%

the name "Loretta" repeated in different contexts and variations

mentions of specific names, particularly those of public figures in legal contexts

No Comments

No Known Activations