INDEX

Explanations

punctuation marks and syntax-related symbols

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

syntactic noise or filler

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

 MainAxisSize

-0.78

 propOrder

-0.67

AxisAlignment

-0.57

 betweenstory

-0.57

RectangleBorder

-0.57

addGap

-0.57

 informée

-0.56

prepareStatement

-0.54

AnchorStyles

-0.53

WaitGroup

-0.53

POSITIVE LOGITS

bestimmungen

0.40

ifiques

0.38

whatever

0.35

 Whatever

0.34

 Autó

0.32

 Anything

0.32

 gydy

0.32

Whatever

0.32

 whatever

0.31

anything

0.31

Activations Density 0.040%

punctuation marks and syntax-related symbols

syntactic noise or filler

No Comments

No Known Activations