INDEX

Explanations

patterns involving numbers and mathematical symbols

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

list items or references

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

 autorytatywna

-0.78

fromnode

-0.71

 betweenstory

-0.63

艋

-0.58

WaitGroup

-0.58

 disambiguazione

-0.56

Aktualisiert

-0.54

WebElementEntity

-0.54

webElementGuid

-0.53

rungsseite

-0.50

POSITIVE LOGITS

Сылтамалар

0.44

 AssemblyVersion

0.41

Nuorodos

0.40

Ārējās

0.38

 Initialized

0.36

DBNull

0.36

nose

0.35

Rohy

0.35

Erreferentziak

0.35

Amour

0.35

Activations Density 0.071%

patterns involving numbers and mathematical symbols

list items or references

No Comments

No Known Activations