INDEX

Explanations

quantitative measurements and time references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 27-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_27/width_16k/average_l0_118

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.27.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

 šaty

-0.55

 dėl

-0.49

 conozco

-0.46

 Kartoffeln

-0.44

 veľa

-0.44

 afbeeldingen

-0.43

 banyak

-0.43

 kebanyakan

-0.43

 pikiran

-0.43

 meeste

-0.42

POSITIVE LOGITS

 ſta

0.65

 faſt

0.63

 ſch

0.62

GEBURTS

0.57

 nanot

0.57

 kilometre

0.57

 ſte

0.56

 metre

0.56

ThroughAttribute

0.55

 checklist

0.54

Activations Density 0.292%

quantitative measurements and time references

No Comments

No Known Activations