INDEX

Explanations

phrases expressing absurdity or frustration

oai_token-act-pair · gpt-4o-mini Triggered by @bot

really neat ideas

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rungsseite

-0.78

 ſind

-0.68

 Autorizaciones

-0.67

 queſta

-0.67

 ſche

-0.64

SceneManagement

-0.63

 незавершена

-0.59

 Exacts

-0.57

 témoig

-0.57

 مرئيه

-0.56

POSITIVE LOGITS

 Knapp

0.35

but

0.35

ln

0.34

 chicos

0.33

</code>

0.33

ba

0.32

 guys

0.32

Zero

0.31

 drauf

0.31

Activations Density 0.037%

phrases expressing absurdity or frustration

really neat ideas

No Comments

No Known Activations