INDEX

Explanations

proper nouns and names associated with organizations or groups

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 doméstico

-0.37

 eletrônico

-0.36

 sledo

-0.35

 tertinggi

-0.34

 ketat

-0.34

 sufrido

-0.33

 teoría

-0.33

 vervangen

-0.33

 gegenwär

-0.33

 vroeger

-0.32

POSITIVE LOGITS

 AssemblyCulture

0.89

parsedMessage

0.82

AndEndTag

0.79

delwed

0.77

 ब्रेकडाउन

0.76

 informée

0.74

ImageContext

0.71

StructEnd

0.70

 noDo

0.67

 Савезне

0.67

Activations Density 0.343%

proper nouns and names associated with organizations or groups

No Comments

No Known Activations