INDEX

Explanations

expressions of truth and authenticity

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

truth and reality

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

anglès

-0.46

 nodding

-0.45

SharedCtor

-0.44

ไว้

-0.44

 inapplicable

-0.43

Приятного

-0.43

 låter

-0.43

 Pastry

-0.42

ʅ

-0.42

ftime

-0.42

POSITIVE LOGITS

本当の

0.65

 REAL

0.64

real

0.60

真实

0.59

 reality

0.58

 réalité

0.58

 realidad

0.58

reality

0.57

 reale

0.57

REAL

0.57

Activations Density 0.038%

expressions of truth and authenticity

truth and reality

No Comments

No Known Activations