INDEX

Explanations

relationships and dependencies between actions and their consequences in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

isNameExpr

-0.35

Kü

-0.35

elif

-0.34

Arn

-0.32

Re

-0.32

rovnik

-0.31

⋀

-0.31

Or

-0.31

Wie

-0.31

POSITIVE LOGITS

 esternos

0.86

parsedMessage

0.86

 EconPapers

0.84

Personendaten

0.83

GOTREF

0.82

ंदीखरीदारी

0.76

 surla

0.73

unknownFields

0.71

EDEFAULT

0.71

principalColumn

0.69

Activations Density 0.275%

relationships and dependencies between actions and their consequences in various contexts

No Comments

No Known Activations