INDEX

Explanations

changes in levels or degrees of various factors, particularly increases and decreases

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

increase and decrease

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

RefNanny

-0.46

IUrlHelper

-0.41

nearest

-0.40

 cistern

-0.40

 Epoch

-0.39

UnusedPrivate

-0.37

 Monks

-0.36

jähriger

-0.36

ArrowToggle

-0.36

outchouc

-0.35

POSITIVE LOGITS

Datuak

0.53

 AssemblyCompany

0.53

</thead>

0.47

íncia

0.46

 HasFactory

0.46

IsMutable

0.45

Sucesor

0.45

 esquecer

0.45

انجليز

0.44

angliski

0.43

Activations Density 0.050%

changes in levels or degrees of various factors, particularly increases and decreases

increase and decrease

No Comments

No Known Activations