INDEX

Explanations

mentions of people holding specific professional positions or job titles

oai_token-act-pair · gpt-3.5-turbo

references to governmental and military positions or service

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

heat

-0.67

iru

-0.66

 fools

-0.66

 onlook

-0.64

 pity

-0.64

 passers

-0.63

goers

-0.63

)].

-0.61

morrow

-0.61

 reproduce

-0.60

POSITIVE LOGITS

 stint

1.07

ionage

0.98

bernatorial

0.95

 retiring

0.87

assador

0.84

utenant

0.84

BuyableInstoreAndOnline

0.83

oslav

0.83

umn

0.81

 overseeing

0.81

Activations Density 0.494%

mentions of people holding specific professional positions or job titles

references to governmental and military positions or service

No Comments

No Known Activations