INDEX

Explanations

mentions of various types of "extensions" related to ladders and other contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

\":

-0.83

-0.79

iard

-0.76

ãĥ£

-0.74

kie

-0.74

ivan

-0.73

icial

-0.73

ky

-0.73

hari

-0.71

INESS

-0.69

POSITIVE LOGITS

 cords

1.07

 cord

0.94

 thereof

0.79

 extensions

0.75

 fuse

0.71

ibility

0.69

 thereto

0.68

 extension

0.67

 grap

0.67

 exerc

0.66

Activations Density 0.023%

mentions of various types of "extensions" related to ladders and other contexts

No Comments

No Known Activations