INDEX

Explanations

references to instructions or specifications within a document

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

start of user input

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

sizeCache

-0.40

fromnode

-0.39

UserScript

-0.38

évaluateur

-0.33

 AppCompatTheme

-0.33

те

-0.32

Yeet

-0.31

出

-0.31

ycznej

-0.31

RegressionTest

-0.30

POSITIVE LOGITS

 AssemblyCulture

0.73

 ivelany

0.67

 kasarigan

0.65

Geplaatst

0.64

BibitemShut

0.59

хьтан

0.58

 désolés

0.54

 queſta

0.54

resizingMask

0.53

bibfield

0.52

Activations Density 0.000%

references to instructions or specifications within a document

start of user input

No Comments

No Known Activations