INDEX

Explanations

The neuron is specifically looking for mentions of different mobile applications within texts

oai_token-act-pair · gpt-3.5-turbo

references to apps

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Memorial

-0.69

 Plains

-0.68

âķĲâķĲ

-0.67

construct

-0.66

stad

-0.66

neck

-0.64

][/

-0.62

 Halls

-0.61

 Baptist

-0.61

bilt

-0.61

POSITIVE LOGITS

 apps

1.14

uggest

0.93

 Apps

0.90

afety

0.86

ointment

0.86

Apps

0.81

alach

0.80

app

0.78

gradation

0.78

 Tinder

0.76

Activations Density 0.011%

The neuron is specifically looking for mentions of different mobile applications within texts

references to apps

No Comments

No Known Activations