INDEX

Explanations

references to structured data or measurements

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

IContainer

-0.59

rungsseite

-0.54

Välislingid

-0.48

 OMITBAD

-0.44

 initComponents

-0.43

 noastre

-0.40

SharedCtor

-0.39

migrationBuilder

-0.38

SBATCH

-0.37

POSITIVE LOGITS

ništ

0.46

WriteBarrier

0.40

 protoimpl

0.38

 wireType

0.38

ethene

0.38

 Grade

0.38

ode

0.37

']>;

0.37

 Interchange

0.37

 Attached

0.36

Activations Density 1.107%

references to structured data or measurements

No Comments

No Known Activations