INDEX

Explanations

terms related to accidents and injuries

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

injuries, accidents, and deaths

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

 resolution

-0.31

 resolutions

-0.30

Resolution

-0.29

ondissement

-0.29

lesias

-0.28

zyden

-0.27

 Resolution

-0.27

Brainz

-0.26

ソ

-0.26

setLayoutParams

-0.26

POSITIVE LOGITS

 الحره

0.58

Autoritní

0.58

 GoogleFonts

0.57

 يتيمه

0.56

 оригіналу

0.56

IVEREF

0.56

osoba

0.52

 mourut

0.52

WriteTagHelper

0.52

 injured

0.50

Activations Density 0.009%

terms related to accidents and injuries

injuries, accidents, and deaths

No Comments

No Known Activations