INDEX

Explanations

terms related to usefulness, convenience, and helpfulness in context

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

useful or helpful tool

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

 ब्रेकडाउन

-0.71

httphttps

-0.57

hyrchwyd

-0.56

NgModule

-0.56

 للمعارف

-0.55

ValueStyle

-0.54

GOTREF

-0.52

 esternos

-0.51

GIVEREF

-0.49

 صوتيه

-0.49

POSITIVE LOGITS

 AssemblyTitle

0.44

 poignée

0.39

 astore

0.39

BES

0.38

 reproducir

0.37

 Италијани

0.37

wark

0.37

gamepad

0.36

 alone

0.36

 empowers

0.36

Activations Density 0.251%

terms related to usefulness, convenience, and helpfulness in context

useful or helpful tool

No Comments

No Known Activations