INDEX

Explanations

references to specific names and significant numerical values

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

academic citations often with year

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

 nahilalakip

-0.52

 especializado

-0.46

 işi

-0.43

 Nachbarn

-0.42

 Usaha

-0.42

 utilidad

-0.42

 CreateTagHelper

-0.41

IsMutable

-0.40

 Chwiliwch

-0.40

 funktioner

-0.40

POSITIVE LOGITS

SerializedName

0.49

setContentView

0.49

requireNonNull

0.48

setVerticalGroup

0.48

invokeLater

0.47

AppMethodBeat

0.46

 gmail

0.46

 defaultstate

0.46

 Official

0.45

Cyfarwyddwr

0.45

Activations Density 0.003%

references to specific names and significant numerical values

academic citations often with year

No Comments

No Known Activations