INDEX

Explanations

proper nouns, particularly names and titles

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

v. or f

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

+:+

-0.56

nodoc

-0.55

+#+#

-0.48

 referenties

-0.48

ponses

-0.46

tagext

-0.45

(!__

-0.44

 transférez

-0.44

Kjelder

-0.44



-0.44

POSITIVE LOGITS

 nahilalakip

0.51

 Chwiliwch

0.50

AccessorTable

0.45

Atentamente

0.44

colgante

0.39

setting

0.38

InjectAttribute

0.38

 ış

0.38

故

0.38

 trekken

0.37

Activations Density 0.199%

proper nouns, particularly names and titles

v. or f

No Comments

No Known Activations