INDEX

Explanations

references to hidden or concealed items

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

items in pockets or clothing

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

 dieß

-0.51

 kommunen

-0.50

RegressionTest

-0.47

 Houſe

-0.47

 myſelf

-0.46

 gospodar

-0.44

 Wikiseite

-0.44

 Polda

-0.41

 Divulgação

-0.41

 ſche

-0.40

POSITIVE LOGITS

AutoField

0.57

IBarButtonItem

0.50

complexContent

0.49



0.47

DockStyle

0.45

 pouch

0.45

 pocket

0.42

XmlAccessType

0.42

SourceChecksum

0.41

UserScript

0.40

Activations Density 0.067%

references to hidden or concealed items

items in pockets or clothing

No Comments

No Known Activations