INDEX

Explanations

specific mention or emphasis on regular procedures or commonly accepted practices

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

asus

-0.72

hung

-0.69

zik

-0.69

rogram

-0.69

gin

-0.69

kamp

-0.68

hani

-0.67

mented

-0.67

haw

-0.67

atoon

-0.65

POSITIVE LOGITS

cy

1.07

ised

1.04

 sized

1.04

ization

1.00

 fare

1.00

 deviation

0.92

isation

0.88

 everyday

0.86

 wisdom

0.85

ized

0.85

Activations Density 1.941%

specific mention or emphasis on regular procedures or commonly accepted practices

No Comments

No Known Activations