INDEX

Explanations

mentions of readers

oai_token-act-pair · gpt-3.5-turbo

terms related to readers and audience engagement

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

 Sabb

-0.68

 Skull

-0.67

Spa

-0.64

--+

-0.63

asia

-0.62

 Winning

-0.62

ayne

-0.62

Alb

-0.61

Rez

-0.60

 Pharma

-0.60

POSITIVE LOGITS

hip

1.63

hips

1.05

uggest

1.04

DragonMagazine

0.88

pace

0.86

boy

0.83

boys

0.81

iences

0.81

="#

0.79

Activations Density 0.030%

mentions of readers

terms related to readers and audience engagement

No Comments

No Known Activations