INDEX

Explanations

technical terms and measurements related to experiments and conditions in scientific contexts

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

âmara

-0.40

Kanpo

-0.36

Externé

-0.36

TokenNameLBRACE

-0.35

 estekak

-0.35

contract

-0.35

toBeTruthy

-0.34

 kaban

-0.34

Tracce

-0.33

 Bíblia

-0.33

POSITIVE LOGITS

:✨

0.57

AlterField

0.48

OrNil

0.48

󠁿

0.47

ChildScrollView

0.46

 ContentValues

0.45

digheden

0.44

inSlope

0.42

 localVar

0.42

 pinulongan

0.42

Activations Density 0.199%

technical terms and measurements related to experiments and conditions in scientific contexts

No Comments

No Known Activations