INDEX

Explanations

quantitative measures and comparisons regarding time frames, measurements, and statistical improvements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 LUMP

-0.40

FormTagHelper

-0.36

󠁬

-0.34

 exactly

-0.34

tanleria

-0.33

rig

-0.32

hells

-0.32

接

-0.32

MLElement

-0.32

 Tempor

-0.32

POSITIVE LOGITS

(>

0.65

Tikang

0.63

 betweenstory

0.56

 PeEnEo

0.54

jewództ

0.54

ніципа

0.52

年以上

0.51

(>

0.50

 NSCoder

0.50

Legături

0.49

Activations Density 0.413%

quantitative measures and comparisons regarding time frames, measurements, and statistical improvements

No Comments

No Known Activations