INDEX

Explanations

code comments in a programming context

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

code comments

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

 sebuah

-0.38

 Sebuah

-0.36

 sacerdotes

-0.35

 mourut

-0.35

 vainqueur

-0.35

AlterField

-0.34

 Bearbeitung

-0.33

 Pertama

-0.33

 barriga

-0.32

 Sicherung

-0.32

POSITIVE LOGITS

 ſind

0.83

AddTagHelper

0.82

 erſt

0.81

tagHelperRunner

0.80

ロウィン

0.79

 للاسماء

0.77

 ་་

0.74

iſen

0.74

Diweddarwch

0.73

 ſehr

0.72

Activations Density 0.445%

code comments in a programming context

code comments

No Comments

No Known Activations