INDEX

Explanations

common second-person pronouns and expressions involving the reader's engagement

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 kasarigan

-0.76



-0.60

 ModelExpression

-0.58

 gyhoeddwyd

-0.56

/**

-0.55

LookAnd

-0.55

 незавершена

-0.54

 AssemblyCulture

-0.53

 Italijanski

-0.52

 Normdatei

-0.51

POSITIVE LOGITS

exitRule

0.54

 équi

0.40

Payload

0.39

EK

0.37

logic

0.37

 Diver

0.37

 Payload

0.37

Sab

0.36

join

0.36

ask

0.36

Activations Density 0.003%

common second-person pronouns and expressions involving the reader's engagement

No Comments

No Known Activations