INDEX

Explanations

phrases related to expressing ideas or opinions in a clear and direct manner

oai_token-act-pair · gpt-3.5-turbo 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

egal

-0.89

lav

-0.70

emis

-0.68

 heavily

-0.65

 particularly

-0.64

ittal

-0.63

 extensively

-0.63

cler

-0.62

orno

-0.61

 defenders

-0.61

POSITIVE LOGITS

 stated

0.79

IFIED

0.73

ify

0.72

 minded

0.70

ified

0.68

 guessed

0.67

 stating

0.65

ername

0.65

ifying

0.64

 clicking

0.63

Activations Density 0.032%

phrases related to expressing ideas or opinions in a clear and direct manner

No Comments

No Known Activations