INDEX

Explanations

code-related syntax involving JSON and object serialization

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

code, test, format, json

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

URIS

-0.42

jsii

-0.41

 saites

-0.39

wel

-0.38

copg

-0.38

ď

-0.37

 pulumi

-0.37

 ویکی‌پدی

-0.37

 arrests

-0.36

owal

-0.36

POSITIVE LOGITS

fjspx

0.67

InjectAttribute

0.54

PerformLayout

0.49

aarrggbb

0.48

новништво

0.46

XtraEditors

0.44

databind

0.44

脚注の使い方

0.43

Vgl

0.42

oa̍t

0.42

Activations Density 0.026%

code-related syntax involving JSON and object serialization

code, test, format, json

No Comments

No Known Activations