INDEX

Explanations

terms indicating collective aspects or comparisons between elements

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

crollView

-0.33

 secund

-0.25

 dispon

-0.25

phosa

-0.24

 folgende

-0.24

έ

-0.24

 couteau

-0.23

 which

-0.23

хівовано

-0.23

POSITIVE LOGITS

owohl

0.74

Personendaten

0.68

ftant

0.66

 Мексичка

0.65

 dieſer

0.63

copg

0.62

 himo

0.61

IsContent

0.60

 genders

0.60

ientras

0.60

Activations Density 0.312%

terms indicating collective aspects or comparisons between elements

No Comments

No Known Activations