INDEX

Explanations

programming-related keywords for including files or libraries

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

code inclusion and requirement

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

 surla

-0.82

 lenker

-0.78

rungsseite

-0.77

ագրություններ

-0.76

 queſta

-0.75

 zuſammen

-0.71

verwijspagina

-0.70

⪜

-0.69

 MainAxisSize

-0.68

<unused79>

-0.68

POSITIVE LOGITS

 calcetines

0.35

adita

0.34

 pantalones

0.33

 vardı

0.33

 penutup

0.33

 panoramique

0.33

include

0.31

 Komponenten

0.31

import

0.30

 Kapitel

0.30

Activations Density 0.001%

programming-related keywords for including files or libraries

code inclusion and requirement

No Comments

No Known Activations