INDEX

Explanations

interconnected phrases within a text

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

sequence for technical or diverse contexts

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

 optimieren

-0.35

 sweet

-0.33

iettivo

-0.31

 aldus

-0.30

weilen

-0.30

 Schönheit

-0.30

 sweetness

-0.29

 sweetest

-0.29

 stripped

-0.28

 SWEET

-0.28

POSITIVE LOGITS

دانشنامهٔ

0.80

MLLoader

0.75



0.73

BeginContext

0.71

contentLoaded

0.71

 يتيمه

0.70

KommentareTeilen

0.69

 nahilalakip

0.69

intios

0.69

fillType

0.68

Activations Density 0.035%

interconnected phrases within a text

sequence for technical or diverse contexts

No Comments

No Known Activations