INDEX

Explanations

specific nouns associated with various categories

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Chwiliwch

-0.51

Personendaten

-0.48

 otomatig

-0.44

 Taktlose

-0.42

GARET

-0.41

ьаж

-0.40

 GoogleFonts

-0.40

 تضيفلها

-0.36

WriteTagHelper

-0.36

 transitional

-0.35

POSITIVE LOGITS

دانشنامهٔ

0.46

CreateModel

0.44

ManagedBean

0.44

ControllerAdvice

0.42

Cast

0.42

Viitteet

0.41

Thing

0.40

 polega

0.40

 Practi

0.40

minimize

0.40

Activations Density 0.009%

specific nouns associated with various categories

No Comments

No Known Activations