INDEX

Explanations

references to physical spaces or areas within a given context

oai_token-act-pair · gpt-3.5-turbo

references to different types of "space" or limitations in space

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

usted

-0.80

ãĥīãĥ©

-0.75

venge

-0.72

ect

-0.70

stal

-0.70

otine

-0.68

inger

-0.67

ument

-0.66

rahim

-0.66

 Simpson

-0.66

POSITIVE LOGITS

 occupied

1.02

 vacated

1.02

 space

0.91

 shuttle

0.90

 spaces

0.90

 Layout

0.90

 occupancy

0.87

 allotted

0.85

 allocated

0.82

 reserved

0.82

Activations Density 0.046%

references to physical spaces or areas within a given context

references to different types of "space" or limitations in space

No Comments

No Known Activations