INDEX

Explanations

various indicators of structure or organization in text

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

${\mkern2mu\underline{\mkern-2mu\smash{something}\mkern2mu }}$

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

 pous

-0.36

Largeur

-0.35

ouin

-0.35

 olduk

-0.35

 Jorg

-0.34

TagMode

-0.33

partements

-0.32

legungen

-0.32

 nasce

-0.32

merksam

-0.32

POSITIVE LOGITS

uxxxx

0.80

 propOrder

0.55

Дереккөздер

0.55

expandindo

0.54

DockStyle

0.54

MigrationBuilder

0.53

elemField

0.49

 تانيه

0.47

#+#

0.44

évaluateur

0.44

Activations Density 0.004%

various indicators of structure or organization in text

${\mkern2mu\underline{\mkern-2mu\smash{something}\mkern2mu }}$

No Comments

No Known Activations