INDEX

Explanations

terms related to legal issues, privacy concerns, and critical evaluations

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

andom

-0.64

chance

-0.63

idency

-0.60

usalem

-0.59

sten

-0.57

insula

-0.56

ighth

-0.56

secution

-0.56

onds

-0.55

riage

-0.55

POSITIVE LOGITS

 enough

1.46

enough

1.12

 Enough

0.99

 insofar

0.99

ly

0.94

looking

0.91

 compared

0.87

isable

0.86

 sounding

0.83

lly

0.82

Activations Density 2.451%

terms related to legal issues, privacy concerns, and critical evaluations

No Comments

No Known Activations