INDEX

Explanations

high numerical values or ten-count thresholds in data

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 11-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_11/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.11.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

RenderAtEndOf

-0.76

 ainfi

-0.73

Viitteet

-0.70



-0.68

transQ

-0.68

 defaultstate

-0.67

SequentialGroup

-0.67

iſen

-0.65

SpringRunner

-0.65

 Geſ

-0.65

POSITIVE LOGITS

[toxicity=0]

0.45

\]

0.42

 поро

0.41

<blockquote>

0.40

</strong>

0.40

ymo

0.39

0.38

OC

0.38

Activations Density 0.024%

high numerical values or ten-count thresholds in data

No Comments

No Known Activations