INDEX

Explanations

terms related to misinformation, misunderstanding, and errors in judgment or representation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/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.21.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

 requis

-0.43

Radius

-0.40

 Currie

-0.40

Required

-0.39

 سور

-0.38

CEP

-0.38

 exigences

-0.38

 yerine

-0.37

 lèvres

-0.37

Uris

-0.36

POSITIVE LOGITS

 wrong

0.75

 falsche

0.68

 WRONG

0.68

 Wrong

0.67

wrong

0.67

 ErrInvalid

0.66

 mistake

0.65

 sbag

0.63

 incorrect

0.63

 misconception

0.62

Activations Density 0.821%

terms related to misinformation, misunderstanding, and errors in judgment or representation

No Comments

No Known Activations