INDEX

Explanations

terms related to exploitation and coercive relationships

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 38-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_38/width_16k/average_l0_128

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.38.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

Rüyada

-0.46

 zdan

-0.41

Và

-0.39

 coinciden

-0.38

 вида

-0.38

 Niemand

-0.37

 soal

-0.37

OrNil

-0.37

 mimo

-0.37

sprechend

-0.36

POSITIVE LOGITS

 ſta

0.60

ſelf

0.59

 pinulongan

0.58

󠁬

0.58

 Anſ

0.56

 vrijwilli

0.55

 ſch

0.55

 hyö

0.54

 raiſ

0.53

ſtra

0.52

Activations Density 0.953%

terms related to exploitation and coercive relationships

No Comments

No Known Activations