INDEX

Explanations

references to old items or concepts

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

old

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 31-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_31/width_131k/average_l0_109

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.31.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

 jinak

-0.45

 einzu

-0.42

inquirer

-0.42

 AssemblyTitle

-0.41

끔

-0.39

 NgModule

-0.39

 MessageBoxIcon

-0.39

 linkovi

-0.38

 darunter

-0.38

 desee

-0.38

POSITIVE LOGITS

 fashioned

0.97

fashioned

0.95

school

0.74

 skool

0.69

ThroughAttribute

0.64

 school

0.61



0.60

didSet

0.54

timers

0.53

 enough

0.53

Activations Density 0.017%

references to old items or concepts

old

No Comments

No Known Activations