INDEX

Explanations

mentions of the word 'weed'

oai_token-act-pair · gpt-3.5-turbo

mentions of "weed," indicating a focus on cannabis-related content

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

ultz

-0.79

 Comfort

-0.72

Riy

-0.68

 Sark

-0.67

vette

-0.66

 Veter

-0.66

 Divinity

-0.63

 Yard

-0.63

ocal

-0.63

 Edison

-0.62

POSITIVE LOGITS

 weed

1.01

killer

1.00

atio

0.91

killers

0.86

weed

0.82

ipop

0.81

 Weed

0.81

 buds

0.75

holes

0.74

bytes

0.74

Activations Density 0.014%

mentions of the word 'weed'

mentions of "weed," indicating a focus on cannabis-related content

No Comments

No Known Activations