INDEX

Explanations

references to errors in machine learning model predictions

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

contentLoaded

-0.49

 tartalomajánló

-0.49

Atsauces

-0.48

 estekak

-0.47

expandindo

-0.47

 שוליים

-0.43

providedIn

-0.43

 kaarangay

-0.43

 setDescription

-0.42

Hentet

-0.40

POSITIVE LOGITS

 attention

0.69

 embedding

0.65

 attentions

0.61

attention

0.60

 Embedding

0.60

 Attention

0.59

Attention

0.58

seq

0.57

 sequence

0.57

 embeddings

0.55

Activations Density 0.268%

references to errors in machine learning model predictions

No Comments

No Known Activations