INDEX

Explanations

references to call events and related procedures

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

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

 Administrativna

-0.50

 AssemblyTitle

-0.48

 name

-0.47

 propOrder

-0.43

ius

-0.41

 named

-0.41

 namorados

-0.41

-0.40

的名字

-0.39

name

-0.39

POSITIVE LOGITS

call

0.71

Call

0.65

 calldata

0.65

 Calls

0.60

CALL

0.59

 Call

0.59

Calls

0.57

 Travail

0.56

posedge

0.56

Calla

0.55

Activations Density 0.015%

references to call events and related procedures

No Comments

No Known Activations