INDEX

Explanations

references to user management and interactions in a system

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

user identity or management

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

ylation

-0.37

美

-0.36

mart

-0.36

jective

-0.36

률

-0.34

言

-0.34

 frivolous

-0.33

初

-0.33

Bag

-0.33

 protest

-0.33

POSITIVE LOGITS

 AssemblyTitle

0.54

 OMITBAD

0.53

httphttps

0.53

 myſelf

0.52

EndContext

0.51

RenderAtEndOf

0.50

rungsseite

0.50

 Conſ

0.49

Identyfik

0.49

 Reſ

0.49

Activations Density 0.008%

references to user management and interactions in a system

user identity or management

No Comments

No Known Activations