INDEX

Explanations

patterns indicating engagement or interaction from respondents

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

HTML or URL fragments

np_acts-logits-general · gemini-2.0-flash

academic citations and code formatting

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

 Hone

-0.73

nteral

-0.71

gdx

-0.67

rouvez

-0.67

 seismo

-0.67

υ

-0.65

ég

-0.65

QM

-0.65



-0.65

③

-0.65

POSITIVE LOGITS

httphttps

1.13

0.87

Somebody

0.81

somebody

0.78

ghurt

0.75

 Somebody

0.75

0.70

0.69

Anybody

0.67

Activations Density 0.881%

patterns indicating engagement or interaction from respondents

HTML or URL fragments

academic citations and code formatting

No Comments

No Known Activations