INDEX

Explanations

terms related to uphill and downhill movements or challenges

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

uphill downhill hills

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

 aanbod

-0.47

elerik

-0.44

 normalidad

-0.44

 coscienza

-0.44

Rao

-0.44

 scienza

-0.43

 poussière

-0.43

 thiệu

-0.43

❚

-0.42

geführt

-0.41

POSITIVE LOGITS

 uphill

2.02

 downhill

1.76

 Upstream

0.70

 upstream

0.65

 upstairs

0.64

 inci

0.64

 hillside

0.63

 hilly

0.60

 upgra

0.60

 offroad

0.60

Activations Density 0.003%

terms related to uphill and downhill movements or challenges

uphill downhill hills

No Comments

No Known Activations