INDEX

Explanations

mentions of asteroids and related terms

oai_token-act-pair · gpt-3.5-turbo

references to asteroids

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

ership

-0.82

ndra

-0.77

ippi

-0.75

ords

-0.74

isdom

-0.72

ORD

-0.72

weeney

-0.71

atche

-0.70

Myth

-0.70

ãĥ¯

-0.69

POSITIVE LOGITS

 asteroids

1.18

 asteroid

0.99

 orbits

0.89

 orbiting

0.88

 habitat

0.81

 circling

0.80

 bombardment

0.78

 habitats

0.78

 dumps

0.77

 Ceres

0.76

Activations Density 0.018%

mentions of asteroids and related terms

references to asteroids

No Comments

No Known Activations