INDEX

Explanations

negative assertions or refusals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 23-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_23/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.23.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

ly

-0.44

 Gegenwart

-0.44

 Unterneh

-0.43

 Bergmann

-0.41

 estudi

-0.40

 Oester

-0.39

IST

-0.38

 Kämp

-0.38

 pasada

-0.37

charAt

-0.36

POSITIVE LOGITS

no

1.04

No

0.97

No

0.91

NO

0.87

principalColumn

0.79

aucun

0.75

 none

0.72

 ſmall

0.72

 nore

0.71

NO

0.71

Activations Density 0.157%

negative assertions or refusals

No Comments

No Known Activations