INDEX

Explanations

insulting language and derogatory terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.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

ValueStyle

-0.59

vated

-0.53

edale

-0.49

 nakalista

-0.47

 Voulez

-0.45

vez

-0.45

query

-0.44

veys

-0.43

URRED

-0.43

irtschaft

-0.43

POSITIVE LOGITS

 assholes

1.01

 bastards

1.00

 bastard

0.99

 asshole

0.98

 fucker

0.97

 shit

0.96

 motherfucker

0.94

 fucks

0.93

fucker

0.92

 bitch

0.92

Activations Density 0.377%

insulting language and derogatory terms

No Comments

No Known Activations