INDEX

Explanations

sections containing code or technical formatting

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 pinulongan

-0.53

AxisAlignment

-0.50

 doubly

-0.49

 Audiodateien

-0.46

(&:

-0.45

↹

-0.44

leyeb

-0.44

ikum

-0.44

řeb

-0.43

cần

-0.43

POSITIVE LOGITS

 '\\;'

0.60

RegressionTest

0.58

ագրություններ

0.58

intios

0.55

 للمعارف

0.54

viewDidLoad

0.54

Quote

0.53

مصادر

0.53

medriver

0.53

ENNIS

0.52

Activations Density 0.155%

sections containing code or technical formatting

No Comments

No Known Activations