INDEX

Explanations

terms related to licensing and permissions

oai_token-act-pair · gpt-3.5-turbo

words associated with licensing and regulation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Bye

-0.62

 bunk

-0.60

 prizes

-0.59

 artisan

-0.57

 transformer

-0.57

 Corpse

-0.56

 Girls

-0.56

 Girl

-0.56

 prize

-0.56

 salad

-0.55

POSITIVE LOGITS

ensing

4.45

ensed

3.28

enses

2.81

ense

2.32

ENSE

2.08

ensation

2.06

ensor

1.87

ENS

1.83

enser

1.69

ensable

1.61

Activations Density 0.013%

terms related to licensing and permissions

words associated with licensing and regulation

No Comments

No Known Activations