INDEX

Explanations

meta information such as dates or comments at the beginnings or ends of articles

oai_token-act-pair · gpt-3.5-turbo

Triggered by @bot

New Auto-Interp

Configuration

jbloom/Gemma-2b-Residual-Stream-SAEs/gemma_2b_blocks.0.hook_resid_post_16384_anthropic

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

HuggingFaceFW/fineweb

Features

16,384

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

HuggingFaceFW/fineweb

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

<bos>

-1.62

<?

-1.32

ⓧ

-1.29

-1.22

/**

-1.14

<?

-1.11

/*

-1.07

/***

-0.97

/*++

-0.81

/*!

-0.80

POSITIVE LOGITS

 maneu

1.41

 affor

1.32

 accla

1.25

 impra

1.21

 increa

1.18

 philanth

1.13

 véhic

1.12

 disagre

1.10

 reluct

1.06

 opport

1.06

Activations Density 0.431%

meta information such as dates or comments at the beginnings or ends of articles

No Known Activations