INDEX

Explanations

phrases related to crime scenes or criminal investigations

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

ename

-0.85

itudes

-0.76

ciating

-0.76

ivism

-0.76

andum

-0.75

rencies

-0.73

ernels

-0.73

anamo

-0.72

iful

-0.71

ahn

-0.69

POSITIVE LOGITS

 closet

1.00

 doors

0.98

 door

0.95

stairs

0.93

 Door

0.92

 balcony

0.91

 bedroom

0.91

 upstairs

0.91

 floor

0.91

 window

0.86

Activations Density 0.027%

phrases related to crime scenes or criminal investigations

No Comments

No Known Activations