INDEX

Explanations

terms related to anonymity and secrecy

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

 Infórmanos

-0.47

पया

-0.43

 einzig

-0.38

 actionMode

-0.37

rowser

-0.35

 bowiem

-0.35

mobili

-0.34

 Producción

-0.34

 Partager

-0.34

andre

-0.33

POSITIVE LOGITS

 mystery

0.77

 anonymity

0.72

 blind

0.72

 Mystery

0.68

 opaque

0.66

 anonymous

0.65

 mysterious

0.64

 shrouded

0.64

AutoScaleMode

0.62

 obscured

0.61

Activations Density 0.285%

terms related to anonymity and secrecy

No Comments

No Known Activations