INDEX

Explanations

instances of the word "to" and its variations, indicating a focus on directional or relational phrases

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

uari

-0.53

 esternos

-0.51

assured

-0.51

vericks

-0.51

INCREF

-0.51

 sortes

-0.49

 fact

-0.48

 quite

-0.48

raska

-0.47

range

-0.47

POSITIVE LOGITS

 greateſt

0.62

 bezeichneter

0.60

 topper

0.59

 quæ

0.58

 toppers

0.57

 Chriftian

0.55

 dovuto

0.54

 infinitive

0.54

 estimés

0.53

 Anſ

0.53

Activations Density 0.185%

instances of the word "to" and its variations, indicating a focus on directional or relational phrases

No Comments

No Known Activations