INDEX

Explanations

references to instructions or requirements in a coding context

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

software development contexts

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

 durumu

-0.43

 Krum

-0.39

ModelAttribute

-0.34

atiche

-0.34

titudine

-0.32

StateMachine

-0.32

 spurred

-0.31

getDoctrine

-0.30

Kjelder

-0.29

 faced

-0.29

POSITIVE LOGITS

MLLoader

0.65



0.58

 pinulongan

0.57

migrationBuilder

0.52

 Италијани

0.51

ſelves

0.49

 Superhosts

0.48

starter

0.48

Carriera

0.48

 Diſ

0.47

Activations Density 0.026%

references to instructions or requirements in a coding context

software development contexts

No Comments

No Known Activations