INDEX

Explanations

references to competitions or selections of individuals or contestants

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

selected individuals or groups

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

KommentareTeilen

-0.44

 resourceCulture

-0.36

postsleuth

-0.36

PMailer

-0.34

 Italijanski

-0.34

้าง

-0.32

 ExecuteAsync

-0.32

AppMethodBeat

-0.32

LookAnd

-0.31

 désolés

-0.31

POSITIVE LOGITS

 selected

0.61

 Selected

0.58

 recruited

0.57

Selected

0.57

 chosen

0.56

selected

0.56

 selects

0.54

InjectMocks

0.54

 trainee

0.52

RTEX

0.50

Activations Density 0.018%

references to competitions or selections of individuals or contestants

selected individuals or groups

No Comments

No Known Activations