INDEX

Explanations

references to variants or variations in a technical or programming context

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

job variants

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

AppMethodBeat

-1.30

 Chwiliwch

-0.98

Chham

-0.95

 BorderSide

-0.91

RectangleBorder

-0.90

 kaarangay

-0.90

 gyhoeddwyd

-0.89

Modific

-0.88

 toJson

-0.85

ValueGenerated

-0.85

POSITIVE LOGITS

job

0.91

mbg

0.89

Job

0.83

job

0.81

Job

0.75

JOB

0.69

JOB

0.69

getJob

0.65

 jobs

0.54

jobs

0.53

Activations Density 0.000%

references to variants or variations in a technical or programming context

job variants

No Comments

No Known Activations