INDEX

Explanations

instances of high numerical values or scores

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 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

MigrationBuilder

-0.69

NUMX

-0.67

 wikipagina

-0.64

mogorov

-0.61

 تانيه

-0.61

rsiniz

-0.60

—

-0.60

SEGUIR

-0.59

Personensuche

-0.59

 Paglinawan

-0.58

POSITIVE LOGITS

0.72

[toxicity=0]

0.59

0.58

*/

0.57

0.56

German

0.55

0.54

0.53

Activations Density 0.882%

instances of high numerical values or scores

No Comments

No Known Activations