INDEX

Explanations

references to popular culture and entertainment

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

Gas

-0.35

 depth

-0.34

 vari

-0.34

PREFERRED

-0.34

 Linux

-0.33

 phase

-0.33

 high

-0.33

 resourceCulture

-0.32

Gaz

-0.32

gas

-0.32

POSITIVE LOGITS

 InputDecoration

0.61

 simplicité

0.60

 banal

0.59

 pragmatic

0.56

 terjangkau

0.55

 nonUne

0.55

 lät

0.53

 simplification

0.53

 jednoduch

0.53

 komer

0.52

Activations Density 0.376%

references to popular culture and entertainment

No Comments

No Known Activations