INDEX

Explanations

specific terms and phrases related to names and historical references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 27-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_27/width_16k/average_l0_118

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.27.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

 laſſen

-0.79

 imagui

-0.78

 otomatig

-0.77

CodeAttribute

-0.77

ſammen

-0.76

 deſſen

-0.76

<unused28>

-0.75

<unused8>

-0.75

<unused3>

-0.75

<unused16>

-0.75

POSITIVE LOGITS

1.96

（

1.37

("

1.27

 (“

1.24

(=

1.20

 (‘

1.14

('

1.13

 („

1.09

(«

1.05

（

1.02

Activations Density 4.686%

specific terms and phrases related to names and historical references

No Comments

No Known Activations