INDEX

Explanations

code elements and terms related to programming and logic structures

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

temporal relationships

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

enderror

-0.80

StoryboardSegue

-0.79

EndContext

-0.78

dafx

-0.75

 CreateTagHelper

-0.74

TestingModule

-0.73

expandindo

-0.72

 surla

-0.68

localctx

-0.68

httphttps

-0.68

POSITIVE LOGITS

 Relaciones

0.39

 później

0.36

 nanti

0.36

 nantinya

0.35

 posteriormente

0.32

 später

0.32

 späteren

0.30

 történ

0.29

 later

0.29

 comparison

0.29

Activations Density 0.063%

code elements and terms related to programming and logic structures

temporal relationships

No Comments

No Known Activations