INDEX

Explanations

syntax related to programming or coding constructs

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

start of turn user

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

Atsauces

-0.51

据

-0.44

crí

-0.41

Erreferentziak

-0.40

 نق

-0.39

 indruk

-0.39

ustainable

-0.38

espoir

-0.38

pter

-0.37

 nadzie

-0.37

POSITIVE LOGITS

setVerticalGroup

0.61

afficheront

0.43

 AspNetCore

0.41

 noDo

0.41

دانشنامهٔ

0.40

最快更新

0.39

featureID

0.39

GOTREF

0.37

 actualización

0.37

querque

0.37

Activations Density 0.014%

syntax related to programming or coding constructs

start of turn user

No Comments

No Known Activations