INDEX

Explanations

web addresses and references to online resources

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

specific capitalized terms

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

 ब्रेकडाउन

-0.68

 kasarigan

-0.64

⟬

-0.57

GTCX

-0.56

 שוליים

-0.51

 reputación

-0.50

 autorytatywna

-0.50

 moeite

-0.49

 touristes

-0.49

 linho

-0.49

POSITIVE LOGITS

oredCriteria

0.34

sy

0.33

FieldBuilder

0.31

sheet

0.30

给他

0.30

 makers

0.30

 Paint

0.30

 нему

0.29

OVID

0.29

 maker

0.28

Activations Density 0.034%

web addresses and references to online resources

specific capitalized terms

No Comments

No Known Activations