INDEX

Explanations

phrases indicating relationships and affiliations

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

цездатний

-0.63

 Савезне

-0.59

posedge

-0.56

 EClass

-0.52

HostException

-0.51

әрмәләр

-0.49

AddWithValue

-0.48

Erfolge

-0.47

 Italijanski

-0.47

лтамалар

-0.47

POSITIVE LOGITS

 various

0.60

 bazı

0.54

 several

0.54

 berbagai

0.53

 some

0.52

 beberapa

0.52

 certain

0.51

 numerous

0.50

 alguno

0.50

 các

0.48

Activations Density 0.604%

phrases indicating relationships and affiliations

No Comments

No Known Activations