INDEX

Explanations

programming-related terminology and constructs within code snippets

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ConstraintMaker

-0.71

onOptions

-0.66

脚注の使い方

-0.61

 ब्रेकडाउन

-0.61

 ویکی‌پدیا

-0.60

KommentareTeilen

-0.59

期刊论文

-0.58

 Wikimedijinoj

-0.56

ofire

-0.56

 kasarigan

-0.56

POSITIVE LOGITS

tagHelperRunner

0.42

 thin

0.40

 indietro

0.38

 adım

0.35

 Präsidenten

0.35

 transmettre

0.34



0.34

einander

0.33

aarrggbb

0.33

 Wikiseite

0.33

Activations Density 0.005%

programming-related terminology and constructs within code snippets

No Comments

No Known Activations