INDEX

Explanations

names of individuals that have been involved in legal or criminal issues

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

LEASE

-0.67

ccording

-0.66

ggle

-0.65

acebook

-0.62

translation

-0.62

ournal

-0.62

LCS

-0.61

 referen

-0.61

 Pradesh

-0.60

Xi

-0.59

POSITIVE LOGITS

Jr

1.37

III

1.22

gren

1.13

Sr

1.10

QC

0.99

aka

0.99

burgh

0.94

quist

0.92

son

0.92

II

0.92

Activations Density 0.824%

names of individuals that have been involved in legal or criminal issues

No Comments

No Known Activations