INDEX

Explanations

phrases that indicate sequencing or methodological processes

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

 SwitchCompat

-0.51

queryInterface

-0.51

 <<<<<<<<<<<<<<

-0.49

GenerationType

-0.45

__(/*!

-0.43

 nahilalakip

-0.43

✭✭

-0.43

 CreateTagHelper

-0.43

Hentet

-0.43

Gizmos

-0.43

POSITIVE LOGITS

Preparación

0.42

HtmlAttribute

0.36

łbym

0.35

UIP

0.35

Eglise

0.34

oughs

0.33

CppCodeGen

0.33

Diámetro

0.33

divarius

0.33

itinéraire

0.33

Activations Density 0.299%

phrases that indicate sequencing or methodological processes

No Comments

No Known Activations