INDEX

Explanations

Twitter retweets

oai_token-act-pair · gpt-3.5-turbo

instances of the word 'Ret' or its variations, likely indicating themes related to retirement or retired individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

STD

-0.72

manship

-0.63

SHIP

-0.62

 Nost

-0.61

 hearts

-0.60

tips

-0.57

 station

-0.56

 sources

-0.56

å·

-0.55

 Passion

-0.55

POSITIVE LOGITS

rieving

1.36

irement

1.33

rieve

1.31

aliation

1.28

ired

1.28

rieved

1.26

ribution

1.23

reat

1.23

ail

1.22

ention

1.19

Activations Density 0.020%

Twitter retweets

instances of the word 'Ret' or its variations, likely indicating themes related to retirement or retired individuals

No Comments

No Known Activations