INDEX

Explanations

This neuron seems to activate on various punctuation marks, symbols, and certain keywords (like "value," "case," "try," "let" and certain short words such as "on", "of", "and") that appear in the contexts of code or legal documents.

oai_token-act-pair · gemini-2.0-flash

&

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-transcoders/layer_25/width_16k/average_l0_41

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.25.ln2.hook_normalized

Architecture

jumprelu_transcoder

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

�

-0.92

Ã

-0.79

&#

-0.72

-0.71

 ‘‘

-0.70

�

-0.70

."<

-0.69

).</

-0.68

¬

-0.63

</

-0.61

POSITIVE LOGITS

2.34

\&

2.14

 $\&$

1.34

&,

0.96

تقاوى

0.90

&.

0.85

＆

0.79

LookAnd

0.78

 एवं

0.76

&[

0.76

Activations Density 8.841%

This neuron seems to activate on various punctuation marks, symbols, and certain keywords (like "value," "case," "try," "let" and certain short words such as "on", "of", "and") that appear in the contexts of code or legal documents.

&

No Comments

No Known Activations

This neuron seems to activate on various punctuation marks, symbols, and certain keywords (like "value," "case," "try," "let" and certain short words such as "on", "of", "and") that appear in the contexts of code or legal documents.

&

No Comments

No Known Activations