INDEX

Explanations

references to graphic or shocking content elements in narratives

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

shortcomings and brutal

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

onOptions

-0.63

culable

-0.62

 パンチラ

-0.60

 deſſen

-0.60

 Menſchen

-0.59

 erſten

-0.59

 CreateTagHelper

-0.59

ſſung

-0.59

 שוליים

-0.58

﹧

-0.58

POSITIVE LOGITS

 oprot

0.45

package

0.35

 Савезне

0.31

 laborales

0.30

<%@

0.30

 bajo

0.29

 líquidos

0.29

 geçir

0.29

 venait

0.29

 sekali

0.28

Activations Density 0.053%

references to graphic or shocking content elements in narratives

shortcomings and brutal

No Comments

No Known Activations