INDEX

Explanations

terms related to various types of attacks or aggressive actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 0-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_0/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.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

 considered

-0.43

 responsible

-0.42

 Mascot

-0.41

ebbe

-0.41

sho

-0.41

 испо

-0.40

 уве

-0.40

 Emails

-0.40

 emails

-0.40

 assigned

-0.40

POSITIVE LOGITS

 attack

1.27

 Attack

1.05

attack

1.02

 attacks

0.99

Attack

0.99

 ATTACK

0.96

 ataque

0.93

ATTACK

0.88

 attaque

0.88

 Attacks

0.87

Activations Density 0.296%

terms related to various types of attacks or aggressive actions

No Comments

No Known Activations