INDEX

Explanations

conjunctions and negation expressions that contrast ideas or elements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_21/width_16k/average_l0_70

Prompts (Dashboard)

36,864 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

WebElementEntity

-0.68

IUrlHelper

-0.61

ніципалі

-0.59

ViewFeatures

-0.59

 незавершена

-0.57

PYX

-0.56

المناصب

-0.54

########.

-0.54

jspb

-0.52

 препратки

-0.52

POSITIVE LOGITS

 exhaustive

0.78

 limited

0.66

exhaus

0.62

0.60

 inclusive

0.56

 keineswegs

0.56

 importantly

0.56

 least

0.55

 exhaus

0.53

 incomplete

0.53

Activations Density 0.145%

conjunctions and negation expressions that contrast ideas or elements

No Comments

No Known Activations