INDEX

Explanations

structured programming elements in code snippets

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 11-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_11/width_16k/average_l0_80

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.11.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

 ―――――

-0.91

 Jefus

-0.90

 myſelf

-0.90

 itſelf

-0.83

ientras

-0.83

 iſt

-0.82

 pageContext

-0.78

Efq

-0.78

 Inscrivez

-0.77

 Monfieur

-0.77

POSITIVE LOGITS

<eos>

0.92

↵↵

0.81

↵↵↵↵

0.67

↵↵↵

0.67

0.65

0.64

">

0.63

</em>

0.63

↵↵↵↵↵

0.63

</

0.63

Activations Density 0.333%

structured programming elements in code snippets

No Comments

No Known Activations