INDEX

Explanations

temporal references and relationships within text

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

final actions

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

RTGC

-0.82

NameInMap

-0.75

 indígen

-0.71

Tembelea

-0.69

PerformLayout

-0.69

 kohdetta

-0.68

 Italijani

-0.68

Ӕ

-0.68

 Италијани

-0.68

BufferException

-0.67

POSITIVE LOGITS

 finally

0.35

principalTable

0.33

Finally

0.32

 Finally

0.31

 handed

0.31

final

0.29

 final

0.29

 furious

0.28

最终

0.28

 handing

0.27

Activations Density 0.014%

temporal references and relationships within text

final actions

No Comments

No Known Activations