INDEX

Explanations

terms related to assessment, evaluation, and the potential for action or improvement

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

 matrimonial

-0.59

 Administrativna

-0.55

INIS

-0.49

ߠ

-0.47

ſammen

-0.47

brunn

-0.46

NewReader

-0.46

菒

-0.45

inales

-0.45

 kernels

-0.45

POSITIVE LOGITS

GOTREF

0.47

 ویکی‌پدی

0.45

 AssemblyCompany

0.44

arraycopy

0.43

hyrchwyd

0.43

parametrize

0.42

setVerticalGroup

0.41

EditorBrowsable

0.41

RTGC

0.39

AddHtmlAttribute

0.37

Activations Density 0.069%

terms related to assessment, evaluation, and the potential for action or improvement

No Comments

No Known Activations