INDEX

Explanations

terms and phrases that indicate specific features, limitations, or implications in a context

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

 otomatig

-0.57

ніципалі

-0.51

 increí

-0.49

 ویکی‌پدیای

-0.49

ervant

-0.47

Personensuche

-0.47

balleur

-0.47

ergies

-0.46

🦵

-0.45

Referanser

-0.44

POSITIVE LOGITS

PerformLayout

0.56

 pinulongan

0.55

 uLocal

0.45

astify

0.38

↵↵

0.38

setVerticalGroup

0.38

 مشين

0.38

0.36

tangentMode

0.36

 invokevirtual

0.36

Activations Density 0.076%

terms and phrases that indicate specific features, limitations, or implications in a context

No Comments

No Known Activations