INDEX

Explanations

references to specific programming structures or classes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.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

 stamp

-0.50

 stamped

-0.48

 kasarigan

-0.48

AutoField

-0.48

ſelf

-0.48

ſelves

-0.45

grès

-0.45

 Stampa

-0.43

#+#

-0.43

'{@

-0.43

POSITIVE LOGITS

 consectetur

1.91

 elit

1.48

elit

0.55

AnimationsModule

0.53

preventDefault

0.43

 tắt

0.41

 crossorigin

0.41

 elites

0.41

 compétences

0.40

IActionResult

0.40

Activations Density 0.000%

references to specific programming structures or classes

No Comments

No Known Activations