INDEX

Explanations

programming constructs and structures like conditions, assertions, and function definitions

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

code context, method calls

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

oa̍t

-0.65

 betweenstory

-0.60

 OMITBAD

-0.58

transQ

-0.56

 noDo

-0.54

évaluateur

-0.54

 ब्रेकडाउन

-0.52

帖最后由

-0.50

 ویکی‌پدی

-0.50

 autorytatywna

-0.49

POSITIVE LOGITS

Personendaten

0.43

ダース

0.42

 similarly

0.40

 equally

0.40

 вида

0.39

 AspNetCore

0.39

fiets

0.38

тари

0.38

 likewise

0.38

 again

0.38

Activations Density 0.084%

programming constructs and structures like conditions, assertions, and function definitions

code context, method calls

No Comments

No Known Activations