INDEX

Explanations

political, ideological, and controversial terms or names

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

<bos>

-2.15

ⓧ

-1.34

/**

-1.22

-1.16

<?

-1.11

<?

-1.06

/*

-1.04

/***

-0.97

///**

-0.95

<!--

-0.81

POSITIVE LOGITS

 affor

1.23

 véhic

1.19

 maneu

1.14

 santiago

1.11

 toledo

1.11

 Minang

1.10

 lidl

1.10

 stockholm

1.09

Juf

1.09

 magis

1.08

Activations Density 0.373%

political, ideological, and controversial terms or names

No Comments

No Known Activations