INDEX

Explanations

proper names related to legal cases or allegations

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

xual

-0.94

UCT

-0.74

BOOK

-0.74

ued

-0.68

ħĭ

-0.67

éĹĺ

-0.65

irens

-0.64

OPLE

-0.63

sembly

-0.62

unct

-0.61

POSITIVE LOGITS

riers

1.04

rier

1.00

rible

1.00

ries

0.99

rigan

0.96

riage

0.96

iage

0.95

ried

0.95

rod

0.92

assment

0.87

Activations Density 1.162%

proper names related to legal cases or allegations

No Comments

No Known Activations