INDEX

Explanations

identifiers, answers, and responses that relate to providing information or clarification

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

question and answer formats

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

tail

-0.38

 lämp

-0.35

oder

-0.35

 personnelle

-0.34

 détermination

-0.34

IContainer

-0.33

 vecind

-0.33

eamiento

-0.33

 éditoriale

-0.32

ed

-0.32

POSITIVE LOGITS

tvguidetime

0.67

SharedCtor

0.63

 ModelExpression

0.63

 EconPapers

0.62

 HasFactory

0.58

 protoimpl

0.58

PMailer

0.55

 Normdatei

0.54

 disambiguazione

0.54

 Italijanski

0.53

Activations Density 0.001%

identifiers, answers, and responses that relate to providing information or clarification

question and answer formats

No Comments

No Known Activations