INDEX

Explanations

technical terms and identifiers in coding or configuration contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 34-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_34/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.34.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

 Spenser

-0.42

 Kelley

-0.36

er

-0.35

 Osborne

-0.35

 Wilde

-0.35

groovy

-0.35

 Мексичка

-0.35

 Ambro

-0.34

 gdyż

-0.34

 Schenk

-0.34

POSITIVE LOGITS

==

0.56

[*]

0.55

 saites

0.53

ahue

0.51

paign

0.51

complish

0.51

isNameExpr

0.50

migrationBuilder

0.50

sement

0.49

 שוליים

0.49

Activations Density 0.753%

technical terms and identifiers in coding or configuration contexts

No Comments

No Known Activations