INDEX

Explanations

phrases related to trending topics or events

oai_token-act-pair · gpt-3.5-turbo

mentions of trending topics and events

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

rem

-0.80

uits

-0.78

 rehe

-0.75

 redress

-0.75

 Squadron

-0.70

 practise

-0.67

 reconnect

-0.66

 sqor

-0.64

 remission

-0.64

 abol

-0.63

POSITIVE LOGITS

Trend

2.32

 Trend

2.26

 Olymp

1.43

Wow

1.40

 pundits

1.35

OEM

1.24

geist

1.17

 Bild

1.11

 pund

1.11

Current

1.07

Activations Density 0.034%

phrases related to trending topics or events

mentions of trending topics and events

No Comments

No Known Activations