INDEX

Explanations

references to dictionary definitions or authoritative explanations

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

dictionary definitions

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

parsedMessage

-0.88

 الرياضيه

-0.81

ագրություններ

-0.72

tinyos

-0.68

 betweenstory

-0.68

 EconPapers

-0.68



-0.67

 ExecuteAsync

-0.67

 تكبرها

-0.66

uxxxx

-0.65

POSITIVE LOGITS

 definition

0.34

 definitions

0.33

 pengantin

0.32

 définition

0.32

 definición

0.31

 definição

0.30

Inter

0.29

line

0.29

 Gesellschaft

0.29

bin

0.29

Activations Density 0.027%

references to dictionary definitions or authoritative explanations

dictionary definitions

No Comments

No Known Activations