INDEX

Explanations

repeated underscore patterns, suggesting it identifies variable names, functions, or data structures in a programming context

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 totalCount

-0.60

httphttps

-0.56

 ſind

-0.55

 faſt

-0.55

 desmotivaciones

-0.55

 ſtand

-0.54

 الحره

-0.54

 iſt

-0.53

 Lainnya

-0.52

 BoxDecoration

-0.52

POSITIVE LOGITS

1.37

\_

1.02

0.80

}_

0.77

‐

0.75

<sub>

0.71

_{

0.71

0.67

$_

0.63

_\

0.63

Activations Density 0.658%

repeated underscore patterns, suggesting it identifies variable names, functions, or data structures in a programming context

No Comments

No Known Activations