INDEX

Explanations

topics related to applied changes and outcomes

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 also

-0.57

 necessaria

-0.48

can

-0.48

 surprisingly

-0.45

 will

-0.44

kste

-0.44

 incidentally

-0.44

There

-0.43

<?

-0.43

 allows

-0.43

POSITIVE LOGITS

 виправивши

0.79

OGND

0.65

GEBURTS

0.59



0.57

PerformLayout

0.53

 typelib

0.52

httphttps

0.51

 चीज़ों

0.50

 Paglinawan

0.49

出版年

0.47

Activations Density 0.029%

topics related to applied changes and outcomes

No Comments

No Known Activations