INDEX

Explanations

relationships between variables in a structured format

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 itself

-0.69

its

-0.66

 яке

-0.59

Its

-0.57

itself

-0.56

 которое

-0.54

Its

-0.53

 său

-0.52

它的

-0.49

its

-0.47

POSITIVE LOGITS

 themselves

1.00

themselves

0.91

 cherchés

0.67

 jotka

0.67

 amelyek

0.66

 которые

0.62

 olduk

0.62

 eivät

0.61

 abstractions

0.60

 generalizations

0.59

Activations Density 2.614%

relationships between variables in a structured format

No Comments

No Known Activations