INDEX

Explanations

numeric information or references in a document

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

+#+

-0.48

 disambiguazione

-0.45

 Füße

-0.44

tembro

-0.44

κτηρισ

-0.43

Viitteet

-0.43

ensdag

-0.43

 gantung

-0.43

Flat

-0.43

 flat

-0.43

POSITIVE LOGITS

LookAnd

0.91

styleUrls

0.37

RenderAtEndOf

0.36

 BorderSide

0.36

 Newt

0.35

RUnlock

0.34

sph

0.34

 CreateTagHelper

0.34

 note

0.33

rrggbb

0.33

Activations Density 0.103%

numeric information or references in a document

No Comments

No Known Activations