INDEX

Explanations

imperative verbs and instructions related to coding actions

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

do stuff here

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

ſſung

-0.69

ScopeManager

-0.69

ロウィン

-0.66

linawan

-0.66

 imagui

-0.65

󠀠

-0.63

ſicht

-0.63

iſchen

-0.63

annica

-0.63

 ſuo

-0.62

POSITIVE LOGITS

 alguno

0.33

OK

0.31

 Meeres

0.29

 opérations

0.29

 alguma

0.29

content

0.28

 elég

0.28

 coś

0.28

model

0.28

相

0.28

Activations Density 0.022%

imperative verbs and instructions related to coding actions

do stuff here

No Comments

No Known Activations