INDEX

Explanations

numbers and numerical data references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ModelExpression

-0.56

TagMode

-0.49

AndEndTag

-0.46

isContained

-0.45

 umur

-0.43

 yaiba

-0.42

хівовано

-0.42

andExpect

-0.41

 dormitórios

-0.40

 νο

-0.40

POSITIVE LOGITS

RegressionTest

0.55

 Signalez

0.48

AnchorStyles

0.47

 bVar

0.46

AutoScaleMode

0.46

'{@

0.45

fromnode

0.42

PreferredItem

0.41

المشاركات

0.40

tdessen

0.39

Activations Density 0.142%

numbers and numerical data references

No Comments

No Known Activations