INDEX

Explanations

code-related terminology and constructs

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

Non-English language related

np_acts-logits-general · gemini-2.0-flash

express intent or action

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

InvalidProtocol

-0.64

 verwijzen

-0.59

########.

-0.57

 يتيمه

-0.57

 sape

-0.57

ierać

-0.56

HttpDelete

-0.56

 ricord

-0.56

 robić

-0.54

rzost

-0.54

POSITIVE LOGITS

RegressionTest

0.67

 worth

0.55

 efficiently

0.53

 somehow

0.50

一度

0.49

 again

0.48

 smoothly

0.48

 سكانية

0.47

 zaak

0.47

 quickly

0.46

Activations Density 0.007%

code-related terminology and constructs

Non-English language related

express intent or action

No Comments

No Known Activations