INDEX

Explanations

phrases that indicate caution or warnings regarding financial or legal responsibilities

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

subject to obligation

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

 ویکی‌پدیا

-0.39

zeniami

-0.38

 Wiktionnaire

-0.38

ถม

-0.38

jähriger

-0.36

 nakalista

-0.36

бражение

-0.35

 ویکی‌پدیای

-0.35

 analysed

-0.35

 různých

-0.34

POSITIVE LOGITS

featureID

0.69

setVerticalGroup

0.60

GOTREF

0.58

WithIOException

0.55

verifyException

0.54

complexContent

0.54

AccessorTable

0.53

AnchorStyles

0.50

uxxxx

0.50

ContentAlignment

0.49

Activations Density 0.034%

phrases that indicate caution or warnings regarding financial or legal responsibilities

subject to obligation

No Comments

No Known Activations