INDEX

Explanations

references to geographical locations, government-related terms like agencies or programs, and incidents of national significance

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Doodle

-0.73

 alloc

-0.62

 preceding

-0.59

 sake

-0.57

 CITY

-0.55

uces

-0.53

Cox

-0.53

 Jackets

-0.53

 Blues

-0.52

 apologies

-0.52

POSITIVE LOGITS

 jeopardy

0.89

fortable

0.88

 danger

0.87

lined

0.82

 trouble

0.82

active

0.81

itialized

0.80

occupied

0.80

oult

0.79

enced

0.79

Activations Density 2.371%

references to geographical locations, government-related terms like agencies or programs, and incidents of national significance

No Comments

No Known Activations