INDEX

Explanations

patterns related to programming syntax and structure

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

special characters in code or data structures

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

Personendaten

-0.57

Попис

-0.54

 autorytatywna

-0.51

 kasarigan

-0.50

 ***!

-0.48

sizeCache

-0.47

 Савезне

-0.46

 Italijanski

-0.45

 otomatig

-0.42

 quæ

-0.41

POSITIVE LOGITS

fjspx

0.55

 Veritas

0.52

 outright

0.51

instan

0.49

ipan

0.49

ynes

0.49

osi

0.49

 also

0.48

 patron

0.47

anson

0.47

Activations Density 0.043%

patterns related to programming syntax and structure

special characters in code or data structures

No Comments

No Known Activations