INDEX

Explanations

words and symbols related to computer code, with secondary activations for survey data, common phrases, and words related to prejudice

oai_token-act-pair · gemini-2.0-flash

New Auto-Interp

Configuration

chanind/gemma-2-2b-batch-topk-matryoshka-saes-w-32k-l0-40/standard/blocks.22.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

32,768

Data Type

float32

Hook Name

blocks.22.hook_resid_post

Architecture

jumprelu

Context Size

1,024

Dataset

chanind/pile-uncopyrighted-gemma-1024-abbrv-1B

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

-1.37

-1.28

↵

-1.23

-1.22

<eos>

-1.19

-1.16

↵↵

-1.16

-1.14

POSITIVE LOGITS

 Theſe

2.33

 Monfieur

2.16

 myſelf

2.13

Efq

2.13

 itſelf

2.05

 Мексичка

1.88

 raiſ

1.79

 purpoſe

1.78

<pad>

1.76

<unused43>

1.76

Activations Density 16.833%

words and symbols related to computer code, with secondary activations for survey data, common phrases, and words related to prejudice

No Comments

No Known Activations