INDEX

Explanations

terms related to temperature, particularly those describing hot

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

hot + noun

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

 betweenstory

-0.66

%)$

-0.57

 للمعارف

-0.55

unce

-0.52

EDEFAULT

-0.50

 Kleid

-0.49

 للاسماء

-0.49

 ſind

-0.49

 Brugge

-0.49

liese

-0.49

POSITIVE LOGITS

Hot

1.30

hot

1.23

Hot

1.22

HOT

1.09

hot

1.08

 hotter

0.96

HOT

0.95

 hottest

0.90

 quente

0.80

 caliente

0.78

Activations Density 0.006%

terms related to temperature, particularly those describing hot

hot + noun

No Comments

No Known Activations