INDEX

Explanations

programming-related terms and structures

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

import model_from_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

zeitige

-0.46

Tasty

-0.44

님

-0.43

BindingResult

-0.43

lusconi

-0.42

lod

-0.41

DKI

-0.41

conexion

-0.40

﹠

-0.40

ientôt

-0.40

POSITIVE LOGITS

 bezeichneter

0.64

adpleegd

0.49

parsedMessage

0.46

 oprot

0.44

 ConstraintSet

0.43

 تضيفلها

0.43

featureID

0.40

 snippetHide

0.40

 kaarangay

0.39

delwed

0.39

Activations Density 0.016%

programming-related terms and structures

import model_from_json

No Comments

No Known Activations