INDEX

Explanations

negations and conditional statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 1-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_1/width_16k/average_l0_69

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.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

+#+#

-0.64

 quelcon

-0.59

 يتيمه

-0.59

be

-0.55

 olsun

-0.54

 contemporaine

-0.50

contentLoaded

-0.50

 africains

-0.49

 ότι

-0.47

 conozco

-0.46

POSITIVE LOGITS

↵

0.52

<bos>

0.50

 Should

0.46

0.45

Should

0.45

0.44

0.43

Gi

0.43

Activations Density 0.096%

negations and conditional statements

No Comments

No Known Activations