INDEX

Explanations

phrases indicating alternative scenarios or possibilities

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

in any case

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

featureID

-0.55

хьтан

-0.55

 queſta

-0.54

 ویکی‌پدی

-0.54

cabulary

-0.54

StreetMap

-0.52

Autoritní

-0.52

osoba

-0.52

Tikang

-0.51

 photolibrary

-0.51

POSITIVE LOGITS

 jedenfalls

0.54

imanapun

0.47

 comunque

0.46

 toekomst

0.41

Ultimately

0.41

theless

0.40

 nonetheless

0.39

 ultimately

0.39

 nikdy

0.39

Nonetheless

0.38

Activations Density 0.020%

phrases indicating alternative scenarios or possibilities

in any case

No Comments

No Known Activations