INDEX

Explanations

numerical or statistical data related to research findings

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

proof headings

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

 vecind

-0.54

 ſta

-0.53

MigrationBuilder

-0.51

StoryboardSegue

-0.51

 campamento

-0.47

RenderAtEndOf

-0.47

Personendaten

-0.46

 boneca

-0.46

 rač

-0.46

JoinTable

-0.45

POSITIVE LOGITS

 nahilalakip

0.53

VolleyError

0.47

yyr

0.44

 становника

0.43

AxisAlignment

0.42

ValueStyle

0.41

zzleHttp

0.41

#%%

0.41

subsubsection

0.39

NonNull

0.38

Activations Density 0.028%

numerical or statistical data related to research findings

proof headings

No Comments

No Known Activations

numerical or statistical data related to research findings

proof headings

No Comments

No Known Activations