INDEX

Explanations

programming-related syntax and structure

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

code, math, and legal contexts

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

 مرئيه

-0.50

 betweenstory

-0.48

 Corée

-0.45

دانشنامهٔ

-0.44

Auszeichnungen

-0.44

 Dumnezeu

-0.43

 volando

-0.43

 mourir

-0.42

 Italij

-0.42

guiente

-0.42

POSITIVE LOGITS

 resourceCulture

1.02

GEBURTSDATUM

0.93

تقاوى

0.89

 autorytatywna

0.88

 lenker

0.86

httphttps

0.84

ngdoc

0.84

SourceChecksum

0.84

awtextra

0.82

 EconPapers

0.81

Activations Density 0.009%

programming-related syntax and structure

code, math, and legal contexts

No Comments

No Known Activations