INDEX

Explanations

words related to crime and legal proceedings

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

è¦ļéĨĴ

-0.86

mma

-0.66

ropolitan

-0.65

lain

-0.64

ording

-0.63

accompan

-0.62

capacity

-0.62

etitive

-0.59

foundation

-0.59

eper

-0.59

POSITIVE LOGITS

Ī

0.80

 into

0.76

coat

0.74

buck

0.74

¸

0.73

ĸ

0.73

awa

0.72

 tides

0.70

 sour

0.70

 Into

0.70

Activations Density 1.498%

words related to crime and legal proceedings

No Comments

No Known Activations