INDEX

Explanations

terms related to technology and potentially dangerous or critical situations

oai_token-act-pair · gpt-3.5-turbo

keywords related to crime, health, and social issues

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

Vox

-0.68

 Galactic

-0.64

 Coin

-0.60

 Planetary

-0.59

 RAND

-0.58

 Slate

-0.58

 Rosenstein

-0.57

Gon

-0.56

[/

-0.56

DSM

-0.56

POSITIVE LOGITS

 mosqu

0.74

alysed

0.73

istries

0.71

fully

0.70

 guiActiveUn

0.67

idently

0.66

aroo

0.66

ullivan

0.66

TPS

0.65

é¾įå¥ĳå£«

0.64

Activations Density 1.240%

terms related to technology and potentially dangerous or critical situations

keywords related to crime, health, and social issues

No Comments

No Known Activations