INDEX

Explanations

instances of the word "last" in various contexts

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

 multiples

-0.33

ture

-0.33

Tener

-0.32

entibus

-0.31

atería

-0.31

radura

-0.31

aqua

-0.30

 Multiply

-0.30

 wireType

-0.30

 Behör

-0.30

POSITIVE LOGITS

 FINAL

0.74

 final

0.71

 posled

0.70

 poslední

0.70

FINAL

0.70

 last

0.67

 Final

0.66

 LAST

0.66

Final

0.65

 letzten

0.65

Activations Density 0.316%

instances of the word "last" in various contexts

No Comments

No Known Activations