INDEX

Explanations

keywords related to programming and variable declarations

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

variable and struct declarations

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

 vielmehr

-0.31

 Versorgung

-0.30

 judíos

-0.27

理

-0.26

 demselben

-0.26

 aikana

-0.26

 laikā

-0.25

 remaja

-0.24

行

-0.24

 ampliamente

-0.23

POSITIVE LOGITS

 Meksiku

0.87

 GenerationType

0.76

ésultats

0.73

 zwiſchen

0.71

Diwedd

0.70

MENAFN

0.69

ſſung

0.69

 imagui

0.69

 Chwiliwch

0.69

niſſe

0.69

Activations Density 0.010%

keywords related to programming and variable declarations

variable and struct declarations

No Comments

No Known Activations