INDEX

Explanations

phrases indicating connections and interactions between people

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

people admired

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

 StringTokenizer

-0.50

帖最后由

-0.46

хьтан

-0.46

 progrès

-0.44

 Desenvolvimento

-0.44

 renseignements

-0.43

 procédures

-0.42

 économies

-0.41

InteropServices

-0.41

 Económica

-0.39

POSITIVE LOGITS

 محفوظة

0.54

 يتيمه

0.50

 оригіналу

0.45

 recons

0.45

AnchorTagHelper

0.44

 somebody

0.42

 anonym

0.40

 Recons

0.40

Somebody

0.40

unt

0.39

Activations Density 0.013%

phrases indicating connections and interactions between people

people admired

No Comments

No Known Activations