INDEX

Explanations

references to professional behavior and etiquette

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

preceding a future tense verb

np_acts-logits-general · gemini-2.0-flash

signals communication

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 24-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_24/width_16k/average_l0_73

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.24.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

ГЛА

-0.48

 salvation

-0.47

thanes

-0.44

holes

-0.44

 Fiche

-0.44

unzel

-0.43

 HttpContext

-0.43

 mecan

-0.43

asi

-0.43

 "}";

-0.43

POSITIVE LOGITS

 signal

1.28

 signals

1.21

 signalling

1.13

 signaling

1.12

 Signals

1.10

 communicates

1.10

 sends

1.08

signal

1.06

 signaled

1.04

signals

1.03

Activations Density 0.286%

references to professional behavior and etiquette

preceding a future tense verb

signals communication

No Comments

No Known Activations