INDEX

Explanations

references to regulatory or operational terms related to the United States Patent Office (USP)

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 paille

-0.46

queles

-0.45

 houten

-0.43

 groote

-0.42

 popol

-0.42

ījum

-0.42

yelidikan

-0.41

 zostać

-0.41

 adicional

-0.41

 Italij

-0.40

POSITIVE LOGITS

USP

2.17

USP

1.52

허

1.06

 resourceCulture

0.83

usp

0.79

SequentialGroup

0.70

RegressionTest

0.66

ValueStyle

0.65

 utafitiHapana

0.63

 informée

0.59

Activations Density 0.000%

references to regulatory or operational terms related to the United States Patent Office (USP)

No Comments

No Known Activations