INDEX

Explanations

references to messages in code or programming contexts

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

message descriptions in code

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

Tikang

-0.63

Diweddarwch

-0.56

 Normdatei

-0.53

SBATCH

-0.52

 MainAxisSize

-0.52

StoryboardSegue

-0.51

IUrlHelper

-0.51

featureID

-0.51

Chham

-0.51

 Infórmanos

-0.50

POSITIVE LOGITS

 déclaration

0.41

 inscrição

0.40

 descrição

0.40

 description

0.40

 déclarations

0.39

 declaração

0.39

 statement

0.38

 descriptive

0.37

 descripción

0.37

 gevoel

0.37

Activations Density 0.002%

references to messages in code or programming contexts

message descriptions in code

No Comments

No Known Activations