INDEX

Explanations

words related to a specific person's name

oai_token-act-pair · gpt-3.5-turbo

mentions of specific names or proper nouns

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

meet

-0.86

ivities

-0.85

prising

-0.84

orative

-0.82

izons

-0.75

opian

-0.75

ivity

-0.74

utterstock

-0.73

innon

-0.73

unin

-0.71

POSITIVE LOGITS

 Ronaldo

0.93

iano

0.85

issance

0.81

zzi

0.79

ivari

0.74

zzo

0.72

ÄŁ

0.68

elli

0.68

ity

0.67

da

0.65

Activations Density 0.022%

words related to a specific person's name

mentions of specific names or proper nouns

No Comments

No Known Activations