INDEX

Explanations

locations and geographical features related to edges and ends of areas

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Wikiseite

-0.34

孔

-0.33

-0.31

̀ng

-0.31

̀nh

-0.30

 venido

-0.30

ąpi

-0.30

 ком

-0.30

ihn

-0.30

POSITIVE LOGITS

extrémité

0.63

 farthest

0.58

 furthest

0.57

 extrême

0.57

WriteBarrier

0.55

extrême

0.54

WithIOException

0.54

tvguidetime

0.54

argout

0.53

 boundary

0.51

Activations Density 0.035%

locations and geographical features related to edges and ends of areas

No Comments

No Known Activations