INDEX

Explanations

URLs and coding-related references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.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

featureID

-0.49

 Exactly

-0.44

 Goal

-0.41

 ddelweddau

-0.40

 Interest

-0.40

лтамалар

-0.40

PMailer

-0.40

 seat

-0.39

RefNanny

-0.39

 esternos

-0.39

POSITIVE LOGITS

wahati

0.43

 вікісторінку

0.42

 ligiloj

0.40

fromnode

0.40

setViewportView

0.40

Izvori

0.39

 MediatR

0.38

__':

0.38

KeySpec

0.36

☚

0.35

Activations Density 0.140%

URLs and coding-related references

No Comments

No Known Activations