INDEX

Explanations

verbs and phrases that indicate improvement or enhancement in processes or situations

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

improve, provide, explain, develop

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

-0.57

-0.38

-0.35

-0.34

。

-0.34

<eos>

-0.33

…

-0.32

“

-0.32

POSITIVE LOGITS

niſſe

0.92

Rüyada

0.91

 ſei

0.90

 ſind

0.89

səhifə

0.88

 ویکی‌پدی

0.87

mpagne

0.86

ſſung

0.85

 ſelbſt

0.84

 שוליים

0.84

Activations Density 0.063%

verbs and phrases that indicate improvement or enhancement in processes or situations

improve, provide, explain, develop

No Comments

No Known Activations