INDEX

Explanations

references to presentations or formal announcements

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

GEBURTSDATUM

-0.58

kijken

-0.41

RenderAtEndOf

-0.40

 velocidad

-0.40

Birch

-0.40

 ancla

-0.39

 velocidade

-0.39

 tari

-0.39

peed

-0.39

BIR

-0.38

POSITIVE LOGITS

 AssemblyCompany

0.65

 attendance

0.50

 Attendance

0.50

bootstrapcdn

0.49

исленность

0.49

GOTREF

0.48

BASELINE

0.48

opsida

0.47

pexpr

0.47

twimg

0.45

Activations Density 0.002%

references to presentations or formal announcements

No Comments

No Known Activations