INDEX

Explanations

phrases and commands related to coding and technical tasks

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

 fubject

-0.63

 pleaſure

-0.59

 Houſe

-0.52

 ſte

-0.48

 للمعارف

-0.48

วะ

-0.47

 houſe

-0.47

parsedMessage

-0.46

 preſent

-0.46

 myſelf

-0.45

POSITIVE LOGITS

 tarik

0.40

aŭ

0.38

 retrospect

0.37

 Nach

0.36

 называ

0.36

turo

0.36

حوالہ

0.36

 careful

0.36

 probed

0.36

opse

0.35

Activations Density 0.950%

phrases and commands related to coding and technical tasks

No Comments

No Known Activations