INDEX

Explanations

coding-related keywords and terms

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

Tikang

-0.43

tanleria

-0.42

 Meksiku

-0.42

 onOptions

-0.41

ArrowToggle

-0.40

 ActionResult

-0.40

￡

-0.39

BagLayout

-0.39

Rhestr

-0.38

ỡng

-0.36

POSITIVE LOGITS

 nahilalakip

0.50

setPreferred

0.43

PreferredItem

0.39

ivores

0.38

SPATH

0.37

 ſon

0.35

艋

0.35

righted

0.35

ouched

0.35

 fubject

0.34

Activations Density 0.186%

coding-related keywords and terms

No Comments

No Known Activations