INDEX

Explanations

programming constructs related to JSON unmarshalling processes

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

code parsing and decoding

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

 campagna

-0.43

Embed

-0.39

couvrez

-0.39

 paix

-0.39

 drept

-0.39

yu

-0.38

 Nigerian

-0.38

Ek

-0.36

enia

-0.36

 Nigeria

-0.35

POSITIVE LOGITS

:✨

0.63

 beginnetje

0.60

 Wicidata

0.58

 wireType

0.54

 يتيمه

0.54

PullParser

0.53

 nakalista

0.52

دانشنامهٔ

0.50

webElementXpaths

0.50

CheckBreak

0.50

Activations Density 0.036%

programming constructs related to JSON unmarshalling processes

code parsing and decoding

No Comments

No Known Activations