INDEX

Explanations

phrases related to custom-made or man-made objects

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "made" and its variations in different contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 retrieval

-0.72

raph

-0.68

 aggregate

-0.65

 hyper

-0.63

 root

-0.62

int

-0.60

shy

-0.60

 tails

-0.59

 flashback

-0.59

 recess

-0.57

POSITIVE LOGITS

made

3.81

Made

2.25

built

1.81

designed

1.62

produced

1.51

crafted

1.39

 Made

1.31

created

1.31

make

1.31

grown

1.30

Activations Density 0.014%

phrases related to custom-made or man-made objects

occurrences of the word "made" and its variations in different contexts

No Comments

No Known Activations