INDEX

Explanations

phrases and features indicative of strong opinions or assertions

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

following sentence starters

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

 molten

-0.35

 disponibilités

-0.32

archiviato

-0.31

ه‌اند

-0.30

akt

-0.30

JSONObject

-0.29

 विश्वसनीयता

-0.29

 пока

-0.28

 Hout

-0.28

ple

-0.28

POSITIVE LOGITS

 Autorizaciones

0.65

balleur

0.62

 AttributeSet

0.54

الحياه

0.54

 متعلقه

0.52

]")]

0.52

LookAnd

0.52

EndProject

0.52

saraba

0.52

AddTagHelper

0.51

Activations Density 0.036%

phrases and features indicative of strong opinions or assertions

following sentence starters

No Comments

No Known Activations