INDEX

Explanations

a specific word related to enjoyment and amusement

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

 Centauri

-0.68

ainer

-0.66

 Lama

-0.65

 defective

-0.64

acid

-0.61

 Journals

-0.61

©¶æ

-0.60

 governs

-0.59

 condemned

-0.58

 condemns

-0.58

POSITIVE LOGITS

nels

1.66

nell

1.23

gal

1.21

nel

1.07

ctory

0.96

icular

0.94

ctor

0.91

eral

0.90

pmwiki

0.88

imation

0.87

Activations Density 0.030%

a specific word related to enjoyment and amusement

No Comments

No Known Activations