INDEX

Explanations

phrases emphasizing totality or universality

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

all/every/always

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

and

-0.39

了呢

-0.36

 somit

-0.32

 then

-0.30

 pinggang

-0.30

 tanggung

-0.29

 clearly

-0.29

 Leder

-0.29

 pack

-0.29

 Krä

-0.29

POSITIVE LOGITS

 betweenstory

0.94

expandindo

0.82

KommentareTeilen

0.82

 AssemblyCompany

0.80

tagHelperRunner

0.79

$_"

0.77

 snippetHide

0.74

ConstraintMaker

0.72

tvguidetime

0.72

ſelves

0.70

Activations Density 0.031%

phrases emphasizing totality or universality

all/every/always

No Comments

No Known Activations