INDEX

Explanations

phrases related to various individuals and proper nouns

oai_token-act-pair · gpt-3.5-turbo

frequent mentions of the word "ra" and references to economic figures and entities

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

kson

-0.71

izabeth

-0.68

ocene

-0.64

 Awakens

-0.63

ĩ

-0.62

 Aval

-0.60

 Papua

-0.60

 Matthews

-0.60

OTOS

-0.59

 Swanson

-0.59

POSITIVE LOGITS

 sidx

0.88

fters

0.87

aca

0.84

eal

0.83

ental

0.81

ught

0.79

asar

0.79

idem

0.78

 fide

0.76

inals

0.76

Activations Density 0.051%

phrases related to various individuals and proper nouns

frequent mentions of the word "ra" and references to economic figures and entities

No Comments

No Known Activations