INDEX

Explanations

phrases related to detailed information, releases, and summaries of various media content such as articles, episodes, and events

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

Downloadha

-0.74

pher

-0.67

 Disciple

-0.66

olson

-0.65

UL

-0.61

among

-0.61

mone

-0.60

duino

-0.59

maid

-0.59

IDS

-0.58

POSITIVE LOGITS

erton

1.00

heartedly

1.00

hearted

0.95

 blown

0.92

 fled

0.92

fledged

0.91

 complement

0.89

screen

0.86

bred

0.83

blown

0.77

Activations Density 3.179%

phrases related to detailed information, releases, and summaries of various media content such as articles, episodes, and events

No Comments

No Known Activations