INDEX

Explanations

mentions involving family members like "grandma," "grandpa," and "niece."

oai_token-act-pair · gpt-3.5-turbo

references to familial relationships and terms of endearment

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ilater

-0.78

 inver

-0.70

nyder

-0.65

 Palest

-0.63

prus

-0.61

 contag

-0.61

 comet

-0.60

gel

-0.59

LW

-0.59

 friction

-0.59

POSITIVE LOGITS

Appearances

0.79

interstitial

0.71

 Riding

0.69

izons

0.68

WithNo

0.67

 sshd

0.66

Â¯

0.64

Grand

0.64

äºº

0.64

////////////////////////////////

0.63

Activations Density 0.095%

mentions involving family members like "grandma," "grandpa," and "niece."

references to familial relationships and terms of endearment

No Comments

No Known Activations