INDEX

Explanations

references to sections within a document

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

section references

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

yntaxException

-0.55

новништво

-0.54

trooper

-0.52



-0.52

exels

-0.52

anyahu

-0.49

 Meksiku

-0.48

♼

-0.48

 torchvision

-0.47

cticut

-0.47

POSITIVE LOGITS

setVerticalGroup

0.49

 propOrder

0.48

 typelib

0.41

قایناق‌لار

0.39

 chapters

0.37

 lộ

0.36

ViewFeatures

0.35

ByPrimaryKey

0.35

kirakan

0.34

 bels

0.34

Activations Density 0.053%

references to sections within a document

section references

No Comments

No Known Activations