INDEX

Explanations

terms related to personal opinions and influences

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

 pleaſure

-0.89

ſelves

-0.84

 propOrder

-0.82

 kasarigan

-0.79

 houſe

-0.78

 ſei

-0.78

 Houſe

-0.77

Personendaten

-0.77

 ſind

-0.76

 ſou

-0.75

POSITIVE LOGITS

0.30

top

0.30

 myself

0.29

我把

0.28

 xấu

0.27

"[

0.27

0.26

[]

0.25

saw

0.25

Activations Density 0.323%

terms related to personal opinions and influences

No Comments

No Known Activations