INDEX

Explanations

code-related elements and constructs within programming languages

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

:+:

-0.43

Rhestr

-0.42

ßerdem

-0.42

 дописавши

-0.40

économ

-0.40

 MonoBehaviour

-0.40

kuuta

-0.38

SPJ

-0.38

 argint

-0.37

 såsom

-0.37

POSITIVE LOGITS

 ComVisible

0.58

 podjela

0.47

 insuffisamment

0.44

 newBuilder

0.43

ActionCreators

0.43

CDS

0.41

vician

0.41

يكب

0.40

 ویکی‌پدیا

0.40

arco

0.38

Activations Density 0.312%

code-related elements and constructs within programming languages

No Comments

No Known Activations