INDEX

Explanations

terms related to gradations or levels of severity and classification in various contexts

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 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

 Absolutely

-0.39

 absolutely

-0.39

 livrer

-0.39

Absolutely

-0.38

Zero

-0.37

eramente

-0.36

 Zero

-0.35

 zero

-0.35

 betreft

-0.35

zeros

-0.35

POSITIVE LOGITS

 autorytatywna

0.69

 moderate

0.69

 medium

0.68

moderate

0.68

ScopeManager

0.68

 Moderate

0.68

 Medium

0.66

medium

0.66

Medium

0.65

Moderate

0.65

Activations Density 0.832%

terms related to gradations or levels of severity and classification in various contexts

No Comments

No Known Activations