INDEX

Explanations

entities, topics, or events that have received high public attention

oai_token-act-pair · gpt-3.5-turbo

expressions related to political controversies and investigations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 exha

-0.77

bes

-0.75

 eleph

-0.74

 newsp

-0.72

 tremend

-0.71

 catast

-0.70

 offending

-0.68

oun

-0.67

 prey

-0.67

 multic

-0.66

POSITIVE LOGITS

Enlarge

1.57

toggle

1.44

Posted

1.35

hello

1.29

Updated

1.26

By

1.20

Guest

1.16

NPR

1.11

Yesterday

1.11

Photo

1.07

Activations Density 0.159%

entities, topics, or events that have received high public attention

expressions related to political controversies and investigations

No Comments

No Known Activations