INDEX

Explanations

instances of potential confusion or ambiguity in text formatting and syntax

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

radiative

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

DeleteBehavior

-0.70

@"/

-0.69

Rüyada

-0.67

IntoConstraints

-0.65

 Normdatei

-0.65

 فريبيس

-0.65

InstrumentedTest

-0.62

sizeCache

-0.60

évaluateur

-0.59

 disambiguazione

-0.59

POSITIVE LOGITS

Tar

0.41

 обо

0.38

RET

0.37

Tar

0.36

KER

0.36

Aware

0.36

shortcuts

0.36

IZON

0.36

Chang

0.35

RED

0.35

Activations Density 0.030%

instances of potential confusion or ambiguity in text formatting and syntax

radiative

No Comments

No Known Activations