INDEX

Explanations

technical or programming-related terms

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

BeginContext

-0.80

 Inscrivez

-0.57

 Shetty

-0.57

tilhe

-0.55

 quæ

-0.55

 queſta

-0.54

 ApiResponse

-0.53

Produkte

-0.52

 dictadura

-0.52

ije

-0.52

POSITIVE LOGITS

al

0.77

Tal

0.60

ral

0.59

 Smal

0.59

ized

0.58

ital

0.58

AL

0.57

Floral

0.57

Jal

0.57

Vital

0.57

Activations Density 1.293%

technical or programming-related terms

No Comments

No Known Activations